Teamwork Graph CLI changelog

The public Teamwork Graph CLI changelog source is:

https://developer.atlassian.com/cloud/twg-cli/changelog.md

This page publishes released sections of the root changelog and derives beta labels from public release manifests.

1.4.4 — 2026-11-05

New Features

• JSM forms and alerts: Discover portal forms, provide readable answers when creating requests, describe, fill, submit, attach, and reopen forms, manage request comments with explicit internal visibility, and triage alerts with actions, notes, and asynchronous request status.
• Confluence: List archived content with --status archived and set, position, upload, or remove page header images without changing the page body.
• Project tags: Replace, add, or remove tags with twg projects update --tags, --add-tags, and --remove-tags.
• Jira replies: Create threaded comment replies with twg jira workitem comment create --parent-id or parentId in --variables-json; preserve reply parent IDs in JSON output.

Bug Fixes

• Login and setup: Fix Windows login freezes and unresponsive cancellation, keep progress output readable, and validate pasted Bitbucket tokens before saving while preserving existing credentials on failure and masking token diagnostics in the terminal.
• Jira embeds: Preserve Confluence whiteboard links as embed cards that Jira can render in descriptions, comments, and transitions.
• JSM request summaries: Create requests with hidden or preset summaries correctly, require a summary only when the request form requires it, and warn when a supplied summary is ignored.
• Skills refresh: Prevent login, authentication refresh, and upkeep from disabling skills; preserve the selected authentication and environment during synchronization. Retain newer installed skills and report a successful cached status without refresh warnings or error exits.
• Search guidance: Quote exact issue keys and error codes in agent search queries while leaving ordinary words unquoted, improving exact-identifier searches.

1.4.3 — 2026-11-02

New Features

• Jira and Trello: Select Jira work items with repeatable --id flags and read Trello card discussions with twg trello card comment query.

Bug Fixes

• Pull requests: Return complete paginated detail collections, filter drafts correctly, and preserve repository names, links, and slug filters when repository lookups are unavailable.
• Jira and Confluence: Reject blank story-point estimates, preserve inline attachment media, correct Confluence browsing links, and retain page-history continuation cursors.
• Validation and connectivity: Reject malformed query dates and nonexistent responsibility dates, honor HTTPS proxy settings during site lookup, and avoid diagnosing network or access failures as expired credentials in twg doctor.
• Skills and agent output: Refresh expiring credentials before skill synchronization, improve cleanup and quarantine failure reporting, and suppress routine upgrade notices in agent mode.

Compatibility notes

• Agent defaults: Set TWG_AGENT_DEFAULTS=1 or pass --mode agent to select agent interaction and output defaults; agent identity detection no longer selects them automatically.
• Whiteboard images: Replace twg confluence content get --format png with twg confluence content export --format png, then call twg confluence content export-status with the task ID on the same site.

1.4.2 — 2026-10-27

Bug Fixes

• Confluence: Markdown content reads preserve links to embedded databases, whiteboards, and pages, including conversion details in --section reads.
• Bitbucket: Repository detection accepts ssh.bitbucket.org remotes and SSH URLs with explicit ports, and skips unusable origins when another valid remote is available.
• Site selection: twg resolve accepts cloud IDs as site selectors.
• Skills: Project-only installation leaves global skills alone, and failed remote skill updates preserve the working local copy.

Enhancements

• Installation: Experimental npm distribution is available for release validation; continue using the hosted installers for supported setup and updates.

1.4.1 — 2026-10-25

Beta release: Available through the beta channel only; not yet generally available.

New Features

• Responsibility activity: Use --include-activity and --since with twg responsibility get, and filter existing candidate groups with --role. Candidate output remains compatible and recorded owners are preserved.
• Confluence groups: Run twg confluence group get --group-id <id> to retrieve group details.

Bug Fixes

• macOS installation: Fix progress displays crashing with EINVAL: invalid argument, kqueue during authentication and the post-login health check. Use static progress status on all Bun platforms to avoid opening an extra terminal input reader.
• Offline skill installation: Complete HTTP retries and fall back to embedded skills when the remote server is unreachable, instead of exiting before installation finishes.
• API-token login: Allow twg login --force --pat to prompt for credentials and a site in agent text mode when a controlling terminal is available.
• Jira input validation: Report conflicting sites in issue ARI batches before attempting organization lookup.
• Bitbucket merge filters: Resolve merge dates recorded as merged-state updates so pull-request queries with merge-date filters return matching results.

Changes carried forward from unreleased 1.4.0

New Features

• Skill refresh: Run twg skills refresh to synchronize authenticated skills across saved profiles, with essential setup, core, Jira, and Confluence guidance available offline before login.

Bug Fixes

• Skill installs and upgrades: Fix refreshes for project-only installs and prevent skill state from newer versions from blocking twg upgrade, while preserving unsupported agent entries, quarantined skills, and cached content when refresh is unavailable.
• Confluence content: Resolve tiny links containing - or _ correctly and cap permissions-list requests at the supported limit of 250, with clearer recovery guidance for unsupported formats and conflicting or incomplete updates.
• Input errors: Report invalid Work and Projects inputs, missing Confluence sites, existing Jira-to-Confluence links, and disabled Code Search apps with specific error codes instead of generic execution failures.
• Connection security: Reject TLS certificate chains containing compromised eMudhra intermediates and refresh trusted certificate authorities in standalone binaries.

Enhancements

• Project date filters: twg projects query now accepts relative dates such as 7d, 2w, 1m, and 1y for --updated-since, --created-since, and --since.

1.3.9 — 2026-10-20

New Features

• Whiteboard PNG exports: Run twg confluence content export <id> --format png, then twg confluence content export-status <task-id> on the same site to retrieve the signed download URL and any export failure details.
• Confluence section reads: Run twg confluence content get <id> --section "<heading>" to read selected sections as Markdown, repeating --section for multiple headings; the new twg confluence page alias also supports content commands.
• Project highlights: twg projects highlights query is now available by default to read project risks and decisions, subject to your project permissions.
• Code search: twg search-code search now includes repository names, code snippets, line numbers, and commands to read the matching revision; code-search commands also accept --repo <workspace>/<repo>.

Bug Fixes

• Bitbucket token recovery: twg doctor and code-search connection diagnostics now recommend twg setup bitbucket --force for invalid or expired tokens instead of a token-generation link with an invalid scope.
• Jira reads: Fix twg jira workitem get targeting the wrong site for issue ARIs, and reject batches containing issues from different Jira sites.
• Search recovery: twg rovo search and twg work search skip failing connectors while continuing to search other sources and retaining usable results and coverage details.
• JSON output: On Linux, agent-mode -o json now produces plain JSON for supported shell file redirects and pipes, including pipes through wrappers; use --output-summary none when a shell replaces itself with the CLI or combines stdout and stderr.
• Confluence content: Preserve table footer rows when writing HTML, and retain more panels, status labels, emoji, and macro information when reading Markdown.
• Reliability: Retry a transient organization-lookup DNS failure once, and preserve successfully retrieved subgraph nodes when other entity types cannot be read with OAuth.

Enhancements

• Command discovery: twg help-map --all expands the full command map, while twg help-map admin expands administration commands from the compact default view.
• Confluence agent output: twg confluence content get -o json inlines page bodies up to 24 KiB per page and 48 KiB per call, with larger bodies saved alongside a section index.
• Antigravity setup: Install skills for Google Antigravity with twg skills install --agent antigravity.

Compatibility notes

• Search exit codes: Skipping unavailable sources during default search can now return exit 0, and other incomplete searches return exit 0 with more than 10 usable results or exit 3 with 10 or fewer; check partial-result markers and failure details instead of treating exit 0 as proof of complete coverage.

1.3.8 — 2026-10-15

New Features

• Profile credentials: Run twg env --exec -- <executable> [args...] to launch a process with the selected profile's credentials, refreshing saved OAuth credentials when due and preserving the child process's input, output, and exit code.
• Integrations: Run twg integrations list to list available integrations.
• Confluence Remix: Create and edit infographics and interactive apps with twg confluence remix infographic create, twg confluence remix infographic edit, twg confluence remix maui create, and twg confluence remix maui edit.

Bug Fixes

• Confluence links: Fix broken page URLs in twg docs query and twg docs get.
• Code review discovery: twg help discover-skills now selects twg-code-review only when you name it explicitly, preserving repository review workflows for other review requests.
• Windows updates: Fix twg update and twg upgrade to run the verified silent installer for current-user MSI installations and keep upgrade notices aligned with the available MSI version.

Enhancements

• Upgrade notifications: Desktop upgrade notifications now identify Atlassian Teamwork Graph CLI and scheduled upkeep, with twg upkeep disable to turn off scheduled upkeep.

1.3.7 — 2026-10-14

Beta release: Available through the beta channel only; not yet generally available.

New Features

• Deployment details: Run twg bitbucket deployment history get to retrieve a deployment's release, commit, pipeline, status, and timing evidence.
• Deployment commit changes: Run twg bitbucket deployment commits query to list commits added or removed since the environment's stored previous successful deployment. Find the UUIDs with twg bitbucket deployment environments query and twg bitbucket deployment history query.
• Commit reports: Run twg bitbucket commit report create to publish a Code Insights report. Reusing the report ID on the same commit replaces that report.
• Confluence approvals: Run twg confluence content approvals pending to list pages awaiting your approval decision.
• Skill management: Run twg skills list to inspect installed skills, twg skills quarantine twg-confluence --yes to disable a skill across upgrades, and twg skills restore twg-confluence --yes to restore it. Replace twg-confluence with the exact skill name you want to manage.
• Loom uploads: Run twg loom video upload to choose visibility and a destination. Visibility accepts owner, workspace, or public; --space and --folder accept IDs, ARIs, and URLs. Space-folder URLs infer their space; space-folder IDs and ARIs require --space. Check space placement in Loom. A failed visibility update exits with code 3 and retains the published video with recovery steps.
• Claude Code transcripts: Run twg session-transcript upload --agent claude-code to upload the transcript named in a Claude Code SessionEnd payload. The previous claude agent name remains accepted.
• Login migration: Run twg auth profile migrate --yes to move older saved logins into profile storage while retaining their names; omit --yes to review the confirmation prompt.

Bug Fixes

• Deployment history: Fix twg bitbucket deployment history query returning HTTP 404, and apply its documented status, pipeline, and date filters.
• Jira team assignments: Fix twg jira workitem update --team rejecting valid Team UUIDs and ARIs when Jira does not enumerate selectable teams.
• Jira batch reads: Preserve issue keys and per-item errors in compact and evidence output from twg jira workitem get --agent-fields @compact or --agent-fields @evidence.
• Rovo errors: Show the upstream failure message and available error codes when twg rovo search fails.
• Usage counts: Include requested GraphQL billing counts for SDK-backed CLI commands such as twg rovo search --usage-counts, in text and JSON output.
• Named skill removal: Fix twg skills uninstall --name to remove only the named skill without disturbing siblings or unrelated agent installations.
• Admin credentials: Honor TWG_ADMIN_TOKEN without reading personal auth storage for admin commands.
• Terminal readability: Render informational text using the terminal's foreground color.
• Installer links: Refresh legacy Statlas installer aliases alongside the canonical macOS/Linux and Windows installers so existing installation links stay current.

Enhancements

• Command planning: twg help describe "jira workitem get" now includes recommended output views and summaries; compact descriptions also identify write access, paid command tiers, and paid options.
• Error recovery: Improve ambiguous-person resolution suggestions and pull-request error guidance with runnable CLI commands and correctly named flags.
• Skill upgrades: Migrate managed skills into shared storage with recovery records and preserved copies, retaining disabled selections and leaving unowned paths unchanged.

Compatibility notes

• Command descriptions: twg help describe now returns compact contracts in a batch envelope by default; run twg help describe "jira workitem get" --detail full for the complete record.
• Older saved logins: If profile selection reports AUTH_MIGRATION_REQUIRED, run twg auth profile migrate. Versions older than 1.3.6 require twg login again after migration.
• Skill agent name: Replace twg skills install --agent rovodev with twg skills install --agent rovo.
• Context retrieval: twg context jira workitem no longer accepts --context-agent-ari; agent-scoped retrieval remains available through the SDK.
• Update manifests: TWG_UPDATE_MANIFEST_URL now accepts only official stable, beta, or versioned candidate manifests under https://teamwork-graph.atlassian.com/cli/. Remove unsupported overrides to use the default manifest, or replace them with an official manifest URL.

1.3.6 — 2026-10-07

New Features

• Native reference support: Pass supported URLs, ARIs, scoped names, and person emails directly to more product commands, with ambiguous or conflicting targets rejected before executing the requested operation.
• Command discovery: Use twg help-map to scan available commands and twg help describe to describe up to 10 commands together, with optional compact output.
• Bitbucket deployment history: Use twg bitbucket deployment history query --environment <uuid> to list deployment attempts for one environment, with status and date filters.
• User memory: Use twg memory create, get, update, delete, and search user-memory in public builds, with --site to select a site.

Bug Fixes

• Jira updates: Fix estimate updates returning empty results and twg jira workitem update --type reporting success without changing the work item type.
• Subgraph exports: Export large subgraphs as JSON, JSONL, text, or GEXF without hitting the limits intended for ASCII graphs.

Enhancements

• Multi-profile login: Create separate organization/environment OAuth or API-token logins with twg login --new-profile, select one with --profile (-p) or TWG_PROFILE, or let --site choose the matching profile; existing logins remain supported. Replace twg env auth, global --env, --stg, and TWG_STG with profile selection (--env remains on login), and replace the former meetings/videos -p with --since and code-search -p with --app. Logout now removes the selected profile and credentials, and rejects --keep-keychain, --keep-file, and --bitbucket; see Manage login profiles for profile management.

Compatibility notes

• Code recall and session transcripts: If you already use twg memory search code or twg session-transcript upload, ask an organization admin to enable agent session transcripts in Atlassian Administration; these operations now require that setting.
• Work portfolios: Replace twg work portfolio --identifier <id> with --identifiers <id> and handle the batch response even for one person; omitting selectors still returns a single current-user portfolio.
• Batch lookups: Split work activity, pull request, portfolio, and people lookups into calls with at most 10 identifiers.

1.3.5 — 2026-10-01

New Features

• Artifacts and Atlas: Delete artifact files with explicit confirmation, and read Atlas project and goal status-update history with pagination and date filters.
• Enriched tools: Add gated tool entries for twg pull-requests get, twg pull-requests query, twg pull-requests search, twg pr-tree, twg work query, and twg workitem-tree.
• Jira batches: Keep failed inputs in ordered Jira work item batch results and return a distinct partial-results exit code for mixed or failed batches.

Bug Fixes

• Site and setup flows: Resolve legacy Jira Cloud and custom-domain sites more reliably across commands and login, and restart Bitbucket token setup automatically when the saved token is invalid or expired.
• Bitbucket pull requests: Accept account IDs and UUIDs in author filters, and fix merged-date filtering.
• Jira work items: Update sprint fields by numeric sprint ID when Jira edit metadata omits allowed values, and report when attachment visibility cannot be determined instead of silently returning none.
• Identity lookup: Match accented and unaccented person names in user-scoped commands.
• Work query: Report unknown PR description completeness as descriptionTruncated: null in full hydration instead of marking upstream excerpts as complete.

Enhancements

• Install and update: Run macOS package upgrades and Windows MSI installs as current-user silent installs where possible, keep package-version notices aligned, and avoid blocking users on versions that are not available for their install source.
• Performance: Speed up Bitbucket merged pull request lookups by overlapping independent requests while preserving existing result coverage.

Compatibility notes

• Raw API requests: twg api has been removed; use scoped CLI commands or SDK tools for supported Atlassian operations.

1.3.4 — 2026-09-29

Beta release: Available through the beta channel only; not yet generally available.

New Features

• Batch reads: Read multiple Confluence pages with twg confluence content get, or multiple Bitbucket PRs with get and activity, with results and failures reported for each item.
• Project highlights and PR windows: Read project risks and decisions with twg projects highlights query, and bound PR merge-date queries with --merged-before.
• Pipeline targets: Use branch, tag, commit, custom pipeline, variable, and --open options consistently across twg bb pipeline run and trigger.

Bug Fixes

• Search and result coverage: Correct combined goal and project filters, incomplete Bitbucket lists, direct-report rosters, collaborator counts, and focus-area tag reads on large sites; retain pagination cursors and partial-result diagnostics in agent output.
• Dates and command reliability: Reject impossible work query dates, accept year-based relative windows, stop queued reads after interruption, and prevent background CLI processes from stealing Windows focus.

Enhancements

• Readable agent output: Keep requested Jira, Bitbucket, and code-search details in compact results, preserve explicit nulls and empty lists, and save large Confluence bodies as readable local files referenced by data.body.outputFile.
• Organization checks: Retry temporary connection failures while keeping access blocked until verification succeeds.

Compatibility notes

• PR date filters: Replace --since in twg pull-requests query and twg bb pull-requests query with --merged-since for merge dates or --updated-since for last-update dates.
• Pipeline starts: Remove the nonfunctional --pr option from scripts, add --custom or --selector when targeting a commit without a branch or tag, and add --open when you need the build URL printed; PR pipeline starts remain unsupported.

1.3.3 — 2026-09-28

New Features

• Bitbucket commit reports: Add twg bitbucket commit report query <commit> to list Code Insights reports for a specific commit.
• Analytics Insights: Add Insights catalog search, get, lineage, SQL query, and AI query commands where available.
• Insights resources: Add twg insights resources create and twg insights resources update for creating catalog entities and updating their documentation metadata.
• Mixed-identity user hierarchies: Query manager chains, direct reports, and peers across Atlassian and connected third-party identities where available.
• Inferred-project metadata: Include title, summary, confidence, activity, and update metadata in twg projects query --include-inferred.
• Pull-request hydration: Add expanded pull-request details and compact reviewer approval totals with pull-requests query --hydrate full.
• Third-party user context: Support connected third-party user identifiers in twg context user for conversation membership and message mentions where available.
• Work query metadata and selectors: Add full work metadata, person selectors, and created, resolved, and merged event selectors to work queries.
• Artifacts: Add artifact file listing, sharing, unsharing, and PNG thumbnail upload support.
• Artifact thumbnails: Add --thumbnail <path> to artifact file create and update commands so PNG previews can be uploaded with artifact content.
• Pull-request mergeability: Add twg bb pull-requests mergeability <id> to report each check's status and whether it currently blocks merging.
• Focus-area tags: Filter focus-area queries by tag and include tag identifiers in query and get results.
• Feedback privacy: Add twg feedback --no-contact so users can opt out while feedback forms default to contact and research consent.
• Plain terminal prompts: Add TWG_PLAIN_TUI for line prompts, numbered menus, and static progress messages in terminals where interactive controls are unreliable.
• Pull request reviewers: Add --reviewer-account-id and --reviewer-uuid to twg bb pull-requests create so reviewer identifier kinds are explicit.

Bug Fixes

• Bitbucket token setup: Fix twg setup bitbucket incorrectly skipping token setup for accounts with Bitbucket access.
• Pull-request diagnostics: Preserve partial-result diagnostics for full pull-request hydration, including query and review failure causes.
• Work search coverage: Retry built-in work sources when connector access fails and preserve available connected-source results with partial-result diagnostics.
• Beta update policy: Fix beta installations being blocked by the stable channel's minimum-version policy, including previously cached policies.
• API site selection: Honor explicit twg api --site selections, reject conflicting selectors, and avoid incorrect saved-site hostnames.
• Subgraph traversal: Fix default twg subgraph queries failing to traverse supported relationships.
• Windows console flashes: Prevent background system lookups from briefly opening a console window on Windows.
• Assets schema settings: Fix Assets object schema settings updates while preserving existing true and false CLI syntax.
• Goal ownership and queries: Fix goal owner assignment and goal-detail retrieval when watcher-access scope is unavailable.
• Jira attachment downloads: Fix attachment downloads when Jira redirects the request to attachment storage.
• API-token access: Fix repeated API-token site-access checks while preserving organization validation and redirect protections.
• macOS installation: Allow the macOS package to install into ~/.local/bin without admin privileges and skip login/setup prompts during current-user installation.
• Upgrade skill refresh: Restore available skills after their availability changes, even when scheduled upkeep is disabled.
• Confluence page reads: Fix reads from viewpage.action?pageId=... links and keep conversion warnings visible in compact output.
• PR tree filters: Fix twg pr-tree merged-date filtering across hierarchy strategies and bound selected-account requests for large organizations.
• Document batches: Avoid repeating the same document activity requests for each ID in a docs get batch.

Enhancements

• Offline capabilities: Keep twg capabilities offline without scheduling background update checks.
• GitHub code-search diagnostics: Explain SAML SSO or private-repository access issues and direct users to reconnect GitHub from Atlassian connected-app settings.
• Readable Jira content: Render Jira descriptions and comments as readable Markdown while preserving JSON output.
• Sprint update guidance: Explain when a sprint cannot be updated through work item editing and provide the board UI workaround.
• Rovo search diagnostics: Expose incomplete searches with exit code 3, keep coverage and failure guidance visible in compact summaries, and explain when retrying will not help.
• Context output: Show names, keys, status, and URLs for directly linked projects in full-detail context and flag generic relationship results limited by --first.
• Hierarchy queries: Improve org-tree, pr-tree, and work-tree performance by using provider-scoped hierarchy traversal by default.
• Bitbucket pipeline output: Keep build status, timing, target, and commit details in compact pipeline output while preserving every build in the full response.

Compatibility notes

• Pull request reviewers: --reviewer is now name-only on twg bb pull-requests create; move Bitbucket UUID values to --reviewer-uuid and Atlassian account IDs to --reviewer-account-id.

1.3.2 — 2026-09-24

Beta release: Available through the beta channel only; not yet generally available.

New Features

• Bitbucket commit reports: Add twg bitbucket commit report query <commit> to list Code Insights reports (CI, security, and code quality) for a specific commit.

• Analytics Insights: Add twg insights resources search to search the Analytics Insights catalog where available, with paging and JSON filters, and twg insights resources get <id> to retrieve full resource metadata. Add twg insights resources lineage <id> to explore upstream and downstream relationships, twg insights query sql to query connected integrations, and twg insights query ai to start or continue AI conversations where available.

• Mixed-identity user hierarchies: Query manager chains, direct reports, and peers across Atlassian and connected third-party identities where available. Accept either identity through --identifier, combine identities with the same email, and avoid duplicate people. Results include actorIdentifiers for follow-up queries. Existing Direct Reports selectors remain supported; queries stop with an error if a hierarchy contains cycles or exceeds supported limits.

• Inferred-project metadata: Include title, summary, confidenceScore, lastActivityDate, and lastUpdated in twg projects query --include-inferred, alongside status and linked evidence. Default agent output includes the metadata and reports its coverage. If full metadata is unavailable, results can still include discovered projects and their supporting evidence. lastUpdated does not necessarily indicate when a project was inferred. Existing scopes, filters, defaults, and projects get are unchanged.

• Pull-request hydration: Add --hydrate full to pull-requests query for expanded pull-request details and compact reviewer approval totals across providers. Missing review evidence is reported explicitly. Existing query defaults and get --full output are preserved.

• Third-party user context: Support connected third-party user identifiers in twg context user for conversation membership and message mentions where available. collaborators continues to accept only Atlassian account IDs.

• Full work metadata: Add Jira project, reporter, creation, and resolution fields, plus Confluence author IDs and page subtypes, to twg work query --hydrate full and twg recently-viewed. Selected records, summary output, and default text rendering are unchanged.

• Work-query person selectors: Add name and email selectors to twg work query, returning the resolved person and any disambiguation information without requiring a separate lookup. Existing identifier queries, activity defaults, and hydration behavior remain unchanged.

• Artifact sharing: Add twg artifacts file share and twg artifacts file unshare to manage access for named recipients, including validating share targets and revoking all explicit user, group, and team grants.

• Artifact listing: Add twg artifacts file list to list your Atlassian Artifacts with cursor pagination.

• Pull-request mergeability: Add twg bb pull-requests mergeability <id> to report each check's type, status, and whether it currently blocks merging. The command does not merge or trigger custom checks. --exclude-git-mergeability-check and --exclude-maximum-commits-behind skip expensive checks but no longer establish full mergeability.

• Focus-area tags: Filter twg focus-areas query --scope all --tag by exact tag name or identifier, and include tag identifiers with --include-tags in query and get results.

• Code memory: Add twg memory search code <query> to search developer memory for the current or an explicitly selected repository where available, with optional branch and file/symbol scopes.

• Work activity: Add created, resolved, and merged event selectors to twg work query. Select by event date, combine selectors without duplicate items, and report matched activities, actor evidence, and coverage in every hydration mode. Existing relationship defaults are preserved. resolved uses the current assignee with resolution history evidence; merged selects authored pull requests by merge date, not the person who merged them. Date-only windows use UTC. Results retain discovered items when full details are unavailable and distinguish incomplete coverage from empty results for each activity and section. Detailed results include full comment text and parent context, plus pull-request repository information and descriptions.

Bug Fixes

• Bitbucket token setup: Fix twg setup bitbucket incorrectly skipping token setup for accounts with Bitbucket access.

• Pull-request partial-result diagnostics: Fix missing partial-result diagnostics in pull-requests query --hydrate full, retaining query and review failure causes in JSON metadata and compact/evidence output.

• Work-search connector access failures: Fix twg work search without a source filter failing on a connector access error by retrying with built-in work sources and reporting reduced coverage.

• Connected-source work search: Fix twg work search omitting results from connected sources. Preserve partial-result diagnostics when a source is unavailable.

• Beta update policy: Fix beta installations being blocked by the stable channel's minimum-version policy, including previously cached policies.

• Explicit API site selection: Fix twg api --site so an explicit site overrides the saved site. Reject conflicting site selectors before sending a request.

• Saved-site hostname resolution: Fix requests targeting an incorrect hostname when saved site settings disagree.

• Default subgraph traversal: Fix default twg subgraph queries failing to traverse supported relationships.

• Windows console flashes: Prevent background system lookups from briefly opening a console window on Windows.

• Assets schema settings: Fix Assets object schema settings updates while preserving the existing true and false CLI syntax.

• Goal ownership: Fix goal creation and updates failing to assign the requested owner correctly.

• Goal query permissions: Fix twg work query --types goals failing to retrieve details when the signed-in account lacks the additional watcher-access scope.

• Jira attachment downloads: Fix attachment downloads failing when Jira redirects the request to attachment storage.

• API-token access: Fix API-token commands failing during repeated site-access checks while retaining same-organization validation and request destination and redirect protections. Explicit access discovery continues to report errors, and attachment downloads and configured remote execution remain supported.

• macOS installation: Allow the macOS PKG to install into ~/.local/bin without administrator privileges, update shell PATH without duplicate entries, and skip login and setup prompts during current-user installation.

• Upgrade skill refresh: Fix twg upgrade failing to restore available skills after their availability changes, even when scheduled upkeep is disabled.

Enhancements

• Offline capabilities: Keep twg capabilities offline without scheduling background update checks.
• GitHub code-search access diagnostics: Explain when twg search-code search results may be incomplete because GitHub requires SAML SSO authorization or private-repository permissions. Direct users to their Atlassian connected-app settings to reconnect GitHub, and omit the generic no-results message when an access warning is present.
• Readable Jira descriptions and comments: Improve Jira work item text output by rendering descriptions and comments as readable Markdown while preserving JSON output.
• Sprint update guidance: Explain when a sprint cannot be updated through work item editing and provide the board UI workaround.

1.3.1 — 2026-09-18

New Features

• People profiles: Describe up to 50 people with repeatable --identifier, --name, or --email selectors, and fetch managers and direct reports together while preserving successful results from partial batches.
• Artifacts: Use Artifacts commands in public builds without feature-flag enrollment.
• Project comments: Read and create Atlas project comments with twg projects comments query and twg projects comments create.
• Terminal completion: Complete shell commands using the local or cached remote tool catalog.

Bug Fixes

• Site targeting: Make twg access list sites with confirmed account activity, block targets outside the saved default site's organization, reject Assets workspace conflicts, and check API-token site access where enabled.
• Work completeness: Prevent silent truncation in large work queries and related context reads, continue through permission-filtered pull-request pages, and preserve partial results with failure details.
• Confluence content: Restore body text in agent evidence and --body-only output, and preserve inline comments, nested formatting, links, and whitespace during HTML conversion.
• macOS login: Fix installer OAuth login failures when the terminal cannot attach the wait spinner.

Enhancements

• Installer integrity: Publish SHA-256 hashes for version-scoped public installer scripts.

Compatibility notes

• People selectors: Replace twg people describe --account-id <id> with --identifier <id>; positional account-ID calls continue to work.
• Default organization: Site targets must belong to the saved default site's organization, so use twg setup default-site --site <site> or twg login to configure an appropriate default before targeting sites in another organization.

1.3.0 — 2026-09-13

New Features

• Work queries: Query work for multiple users with repeatable --identifier options, with third-party identities and linked same-email identities supported where enabled.
• Manager discovery: Find managers using third-party identities where enabled, while retaining existing Atlassian account inputs.

Bug Fixes

• Command completion: Fix foreground CLI commands that remain running after their work completes.

1.2.9 — 2026-09-09

Beta release: Available through the beta channel only; not yet generally available.

New Features

• Confluence space access: Add commands for space permissions, roles, role assignments, and custom-role management, including role-aware space creation and JSONL output.
• Cross-product identity queries: Support feature-gated third-party user identifiers and bounded actor scopes for Docs, Calendar meeting, and video queries, including same-email identity relationships where available.
• Workspace administration: Add --include-permissions to workspace-member queries for authorized permission audits.
• Work search completeness: Return exit code 3 when twg work search has incomplete source coverage while preserving available results, partial markers, and structured failure details.
• Project-scoped skills: Add twg skills install --local to install skills in the current project rather than the shared agent home.
• Assets discovery: Accept twg assets schema query as an alias for schema listing and clarify the schema-ID discovery flow in command help.
• Code-review workflow: Include the public twg-code-review skill and its artifact helper in public builds.

Bug Fixes

• Confluence content reads: Read labels from folders, whiteboards, databases, and other supported content types; keep pagination cursors out of prose; and materialize large Maui source files instead of printing them inline by default.
• Confluence HTML conversion: Prevent duplicate formatting marks while preserving repeatable annotations and innermost formatting attributes.
• Work and project queries: Correct third-party pull-request and document counts, and resolve named users for twg projects query --scope user so inferred projects are included.
• Bitbucket pull requests: Return canonical repository URLs for ARI inputs, report draft state consistently, and provide explicit --workspace and --repo retry forms when context cannot be inferred.
• Administration output: Show complete user details in twg admin user list text output and keep unavailable Confluence content-status listings hidden from SDK hosts that cannot serve them.
• Uninstall targeting: Prevent Linux uninstall planning from treating an unrelated /usr/local/bin/twg as a macOS package installation.

Enhancements

• HTTP transport: Move source and standalone builds to Bun 1.4.1 with native HTTP/2 fetch and HTTP/1.1 fallback, replacing the custom pooled HTTP/2 transport for more reliable requests.
• Work-tree queries: Report the effective activity window and default activity counts to the last 30 days to avoid accidental unbounded graph queries.
• Installation and cleanup: Make setup verification clearer, attribute installs launched from supported IDE integrations, and preserve coexisting or unsafe installation files during uninstall rather than following or deleting them.
• Authentication recovery: Provide clearer guidance when a modern scoped API token cannot read a requested field.
• Grok Build: Add native plugin metadata, installation guidance, and installer attribution support.
• Output and guidance: Keep Confluence pagination cursors separate from text output, explain invalid-cursor recovery, and direct Bitbucket deployment-history reads to the deployment command.

Compatibility notes

• Work-tree activity: Activity counts now default to the last 30 days; pass --since all when intentionally requesting an unbounded history.
• Docs user selectors: Use --identifier for new Docs query and get integrations; existing --account-id scripts continue to run through a hidden compatibility option.

1.2.8 — 2026-09-06

New Features

• Bitbucket workflows: Query draft pull requests and deployments, create repositories, inspect or update Bitbucket Pipelines configuration, and list file-level pull-request merge conflicts before merging.
• Confluence workflows: Manage approvals and content status, list available statuses, check content permissions, fetch AI summaries, inspect date-bounded analytics with agent viewers, and update whiteboards or databases using exact context files.
• CLI output, search, and authentication: Add bounded Jira field projection, evidence metadata, direct JSON/JSONL output control, billed usage totals, Feedback App filters by JPD idea or space, feature-gated Rovo deep search in public builds, third-party identity fallback, and encrypted credential storage with bidirectional migration, gated raw inspection, and connectivity diagnostics.

Bug Fixes

• Queries, mutations, and structured output: Resolve Bitbucket repository slugs and named-person pr-tree roots, Jira parents, Rovo Slack results, inferred projects, mixed-case cloud IDs, and linked third-party identities correctly; repair meetings queries and limits; reliably deliver Jira transition comments; return actionable Confluence edit and Remix errors; advertise the actual work-tree JSON fields; and emit nested primary collections correctly in JSONL.
• Authentication, installation, connectivity, and embedded execution: Reuse stored OAuth for twg login --no-verify and remote help, accept server catalog grammar, skip installer login in headless environments, avoid unreliable pooled update checks, prevent captured Windows version output from hanging, isolate timed-out just-bash executions, make credential migration race-safe, preserve Linux scheduler definitions and data when removal fails, remove complete Windows and copied skill installations, and classify transport handshake failures as retryable network errors.
• Help and billing accuracy: Stop advertising unsupported pull-request, Jira, and Assets options, and correct Assets, responsibility, user, and people billing attribution and feature gating.

Enhancements

• Coverage and enriched metadata: Report Confluence search and Jira hydration coverage, expose embed and Smart Link destination URLs, show whether a Confluence space has instructions, label paid command paths consistently, include field-level usage details and totals in billing metadata, and send only the current Atl-Ubp-* billing headers.
• Updates: Show clearer transfer progress and recover more reliably from stalled or failed downloads.

Compatibility notes

• Inferred teams: Update stored inferred-team identifiers to use first-class InferredTeam workspace ARIs; legacy masked IDs are no longer accepted by inferred-team queries or follow-up commands.

1.2.7 — 2026-08-27

New Features

• Bitbucket: Add deployment-environment queries, workspace-member lookup, repository directory browsing, pull-request diff anchor resolution, and reviewer selection by Atlassian account ID or Bitbucket UUID.
• Confluence: Add commands for content-approval state, per-user content analytics, and Maui remix retrieval.
• Installers and local management: Add signed Windows MSI installers and explicit twg uninstall selectors for all, direct, and macOS package installations.
• CLI vault: Add opt-in encrypted credential storage with OS-keychain-backed keys and recovery and administration commands, while preserving existing auth.conf installations.

Bug Fixes

• Queries and targeting: Fix Confluence space search, dummy-cloud routing, natural-language organization targeting, and startup when SharedArrayBuffer is unavailable.
• Partial results: Preserve usable work query and docs query rows when later GraphStore pages fail, expose diagnostics, and return exit code 3 for incomplete results.
• Structured output and Confluence access: Fix JSON/JSONL error envelopes, permission normalization, and output-file truncation warnings.

Enhancements

• Authoring guidance: Clarify Confluence layout, access, and visual-aid guidance for more predictable page creation and updates.

Compatibility notes

• Bitbucket deployments: Use twg bb deployment environments query instead of the renamed twg bb deployment query command.

1.2.6 — 2026-08-23

New Features

• Confluence content status: Add twg confluence content get-status for pages, blog posts, and live docs.
• Confluence macro resolution: Add asynchronous content macros resolve and resolve-status commands for converting supported macros to HTML.
• Bitbucket deployments and reviews: Add deployment lookup by UUID, multi-line inline comment and suggestion ranges, and --pending comments for unpublished review feedback.
• Jira work item reads: Add --remote-links and --batch-concurrency to twg jira workitem get; --full includes comments and remote links, repeated reads are batched, bulk reads are chunked, and transient rate limits are retried.
• Jira Service Management participants: Add request-participant list, add, and remove commands with me, email, and Atlassian account ID selectors; non-interactive removal requires --yes.
• People and project context: Add name and email selectors to twg collaborators, Atlassian Project update notes to twg projects get, and positional issue identifiers for JSM incident and post-incident-review reads.
• Pipeline filtering: Add --pattern to Bitbucket pipeline query and latest-failure commands for filtering custom pipeline selectors.
• Rovo connector authorization: Expose the validated connector authorizationUrl from twg rovo auth <app> --output json without opening a browser.

Bug Fixes

• Confluence content fidelity: Preserve table widths, image dimensions, breakout widths, inline media types, smart links, attachment chips, panel tables, and emoji normalization through ADF and HTML conversion.
• Confluence routing and creation: Keep URL-based commands on the tenant embedded in the URL, fix synced-folder creation, and return correct Confluence page URLs from docs query.
• Context hydration: Fix twg context get --detail full so Jira and Confluence anchors and relationships load with dates, status, and URLs when generic relationship discovery is unavailable.
• Document pagination: Prevent duplicate documents across pages, retry stalled cursors, terminate cursor cycles, lower unsafe automatic page sizes, and return resumable partial results with exit code 3 instead of discarding completed work.
• Document identity and counts: Reconcile duplicate third-party document records, include ownership and collaboration activity from connected accounts, deduplicate logical documents by URL, combine activity tags, and align counts with returned items.
• Work and organization results: Include code, design, calendar, and document activity from connected accounts; fix inferred Atlassian Project discovery and totals; preserve fuzzy goal results; and prevent work-tree identifier-resolution failures.
• Work-tree resilience: Retry narrower report batches after backend count-request failures and return unaffected result categories with partial-result warnings when one category remains unavailable.
• Site targeting and login: Honor explicit --site and bare-UUID --cloud-id selection, including schema-driven commands; reject conflicting selectors, prevent stale cloud IDs from crossing tenants, and avoid malformed query-context headers.
• Structured and piped output: Prevent silent stdout truncation on non-blocking pipes, keep YAML validation failures in YAML, preserve Search Code scan cursors, normalize live symbol responses, and retain metadata fields in agent-mode output.
• Command correctness: Fix connector-filtered Rovo search, Bitbucket merge waiting, Jira Service Management approval completion output and query aliases, and user-search --name me resolution.
• Runtime reliability: Stop detached telemetry workers after their parent exits so they cannot survive indefinitely and consume CPU.

Enhancements

• Confluence comments: Move comment commands to cursor-based APIs with ok, comment or comments, nextCursor, failure guidance, optimistic --expected-version, complete filters, the md body-format alias, and optional reply expansion through --include-replies true.
• Work detail levels: Add explicit none, summary, and full detail levels and load selected entity details from supported services.
• Update channels: Report channel and install-channel metadata, surface channel-mismatch and downgrade-available states, support confirmed moves from beta to an older stable build, and include redacted failure stage and install-method diagnostics.
• Richer Confluence reads: Accept positional identifiers in confluence space get, return space URLs, enabled content types, and instructions, and show total content views in metadata and text output.
• CLI compatibility and safety: Restore Jira Service Management list and --limit aliases and preserve explicit confirmation before destructive commands run.

Compatibility notes

• Confluence comments: Remove legacy create --inline and --content-type, resolve/reopen --type, list --inline and --type, --comment-type all, auto-pagination, and --limit 0; list pagination now accepts limits from 1 to 250, omitted --comment-type returns footer and inline comments, and replies are excluded by default; pass --include-replies true to include them.
• Required space context: twg confluence content list now requires --space.
• Bitbucket pipelines: pipeline run and pipeline trigger now return after submission; use pipeline get to inspect a run or pipeline tail to follow logs because the standalone pipeline wait command and create-time wait flags have been removed.
• Search Code cache controls: Remove the obsolete Search Code cache environment variables and cache-clear flags.

1.2.5 — 2026-08-13

New features

• Add --content-width to Confluence content create and update commands for pages, live docs, and blog posts.
• Add --synced-container-url to twg confluence content create so a new folder can be linked to an external Google Drive or Dropbox folder and synced. The command reports the linked source and sync status.
• Add dedicated Jira Service Management Help Center query commands, replacing the former combined Help Center query command.
• Add YAML output with --output yaml or --output yml. YAML uses the same response envelope as JSON and supports YAML output-summary files. Follow-mode commands buffer their output before emitting YAML.

Bug fixes

• Confluence HTML conversion now handles inline code and nested code blocks correctly, preserves media attribution metadata, and renders infographic media at the intended width.
• Confluence public-link requests now reject unsupported content types with a clear message instead of attempting a page-only operation.
• Confluence search results now link directly to the relevant site pages.
• First-time Confluence reactions are now reported as added.
• Assets object updates and deletes now accept UUID reference IDs.
• Windows console flows for login, Bitbucket setup, skill installation, uninstall, and consent now work with typed prompts and cleaner PowerShell output.
• Unsupported Confluence status updates now return a failing exit code.
• Numeric Confluence references now resolve correctly.
• Input validation errors now remain distinct from backend response errors and no longer include unrelated diagnostics.
• Search warnings now show a clear user-facing message instead of exposing a service URL, while partial results remain available when possible.
• twg search-code file now honors --output-file, renders source URLs once, and gives provider guidance only for the selected source.
• Invalid twg search-code queries and removed commands now return clear usage guidance.
• Jira duplicate detection now searches both older and newer work items by default.
• Zero-length relative durations such as 0d and 0w are rejected with a clear validation message.
• Invalid draft requests for unsupported Confluence content types now return a clear error.
• Time-bounded relationship queries now return correctly ordered results and clearer errors when a pattern is unsupported.

Enhancements

• Fixed-value options now accept case and separator variants, such as at-risk, AT_RISK, and at_risk.
• Confluence guidance now makes it clearer when to create a live doc versus a classic page.
• Confluence content create and update errors now include the reason for failure and suggested next steps.
• Agent-mode search responses now preserve warnings and failure details for zero-result and partial searches.
• Agent-mode comment and task resolution responses now include the complete mutation result.
• Meetings lookups now use one fallback search per request and combine not-found messages for easier troubleshooting.
• User selectors now accept me and currentUser to refer to the authenticated user.
• Projects, goals, and focus areas now accept relative time filters such as 7d, 2w, and 1m.
• The Jira work-item comment query example now uses descending comment order.
• The Confluence search --body compatibility alias has been removed; use --excerpt for highlighted excerpts.

1.2.4 — 2026-08-07

Beta release: Available through the beta channel only; not yet generally available.

Minor Changes

• 49def8e: Expose twg assets graph schema and twg assets graph connections.

• e7b36eb: Add formatted transition comments to twg jira workitem transition using the same --transition-comment and --transition-comment-format contract as twg jira workitem update.

  --transition-comment remains optional. When supplied, the command must also specify its transition target (--status for update or --transition-id for transition). Supported formats are html, markdown, and plain, with html as the default.

• e0b6422: Add --scope org support to twg projects query --include-inferred, returning Atlassian Projects and inferred projects across resolved reporting chains.

• 980c6ec: Migrate context jira-workitem, context confluence-page, context confluence-space, context confluence-blogpost, context confluence-whiteboard, and context user commands to the @atlassian/twg-tools SDK. All context subcommands are now isolate-safe and host-neutral. CLI behavior is preserved byte-for-byte. Old app-native projection files removed.

Patch Changes

• d4d3b85: Fix five twg confluence space defects:

  • space get / space list now request description-format=plain, so a space description is returned instead of always being null.
  • space get --labels / space list --labels now send include-labels=true. The flag was previously read for rendering only and never reached the API, so it could never return labels.
  • space create / update / archive / unarchive / delete write their confirmation preview to stderr instead of stdout, so -o json output stays parseable when --yes is omitted.
  • space list --limit now rejects values that are not positive integers instead of silently coercing them (abc became 25, 0 and -5 became 1). The rule lives in normalizeConfluenceSpaceListArguments via the exported parseConfluenceSpacesListLimit, so executeCmd / executeTool callers get it too and not just the native CLI. Values above the API maximum still clamp to 250 with a warning.
  • space create's summary no longer advertises a global/personal --type option that the command does not accept; it documents --private instead.

• a4d1b85: Fix jsm form answers list to return every form question without relying on formatted answers field keys.

• 23c9ecd: Restore user bulk-lookup email and legacy-username lookups without an explicit --site by preserving the site auto-loaded from CLI authentication. Account-ID-only batches continue to use the site-less GraphQL path.

1.2.3 — 2026-08-07

Beta release: Available through the beta channel only; not yet generally available.

Highlights

• JSM forms: add commands to list and retrieve forms attached to work items, inspect normalized field metadata, and return formatted form answers.
• Confluence ownership and content: add twg confluence content set-owner for transferring page ownership, correctly resolve embed and smart-link URLs, and restore --body-file and --edits-file updates.
• Confluence tasks: automatically retry complete and reopen operations when concurrent page changes cause transient conflicts.
• Team discovery: add --include-inferred to team queries, team reads, and member lists so inferred teams can be included alongside canonical teams.
• Cross-product spaces: twg spaces query|get -o jsonl now returns JSON Lines instead of silently falling back to text output.

Compatibility notes

• Video transcripts: twg videos get --transcript now requires none, preview, or full; replace --transcript-output-file with --transcript full --transcript-only --output-file <path>.

1.2.2 — 2026-08-05

Beta release: Available through the beta channel only; not yet generally available.

Highlights

• Unattended installation: public installs using --yes --skip-login --skip-skills now carry explicit consent through finalization, restoring fully headless RovoClaw-style installation without piped prompt input.
• Projects: twg projects query --scope me --include-inferred can include active projects inferred from linked Jira and Confluence work, with source labels and coverage warnings.
• Confluence: add twg confluence content convert to switch content between page and live-doc modes.
• Goals: status-update creation and updates now accept the documented target-date confidence values and reject invalid scores clearly.
• Portable command coverage: Assets type queries, Bitbucket read commands, and Subgraph commands now run through the host-neutral SDK while preserving their CLI behavior.

1.2.1 — 2026-08-03

Beta release: Available through the beta channel only; not yet generally available.

Highlights

This release includes user-facing changes merged after 1.2.0.

• 1.2.0 command changes: the complete public command comparison tables are retained in the 1.2.0 release notes, where those syntax updates shipped.
• Confluence comments and stars: add content and space star/unstar commands; comment lists now include footer and inline comments by default, and comment replies and updates correctly render Markdown and accept body files.
• Confluence content lifecycle and output: add explicit trash, restore, and permanent-purge operations; HTML returned by content and version reads is now formatted for readability.
• Goals: add twg goals create, twg goals update, and twg goals archive.
• Goal status updates: add twg goals status-update create and twg goals status-update update for posting and revising goal updates.
• Bitbucket commits: add twg bitbucket commit create to create add, update, or delete commits remotely without a local clone, with an explicit target branch and automation safeguards.
• Pull-request data: twg pull-requests get, query, and full pr-tree JSON output now include reviewers and accurate authenticated-user role tags.
• Clearer search and time windows: user searches with no matches provide actionable guidance; work results include the matched activity timestamp; and notifications support --since, --from, and --to.
• Confluence analytics: content analytics now return structured failure reasons when an operation cannot complete.
• Loom video library: add commands for comments, reactions, generated content, transcript retrieval, uploads, sharing and permissions, recovery, and folder creation; video and comment output now resolves owner and author identities.
• More dependable automation: add --select <paths> to project command output, improve structured failure recovery details and option repair guidance, and include browsable URLs in Jira work-item responses when a site URL is available.
• Trello and Subgraph compatibility: card-description updates accept 8-character Trello shortlinks, and twg subgraph again supports the advertised ascii and gexf output formats.

Compatibility notes

• Confluence comment lists: twg confluence content comments list now defaults to --comment-type all, so --limit applies separately to footer and inline root comments. Pass --comment-type footer to retain the former footer-only result set.

1.2.0 — 2026-08-02

Beta release: Available through the beta channel only; not yet generally available.

Highlights

• Safer automation: noninteractive consent and destructive actions now require explicit --agree or --yes, avoiding unexpected prompts in agent, JSON, and no-terminal runs.
• Confluence workflows: content and space addressing are more consistent; you can watch or unwatch content, spaces, and labels; space list accepts --key and caps --limit at 250; and task, comment, search, CSV update, and space-instruction handling is more reliable.
• Jira productivity: discover editable built-in fields with --include-system-fields, set a reporter when creating work items, find similar work items in public builds, create issue links in the requested direction, and receive stable JSM SLA cloud routing metadata in JSON output.
• Team and identity management: add or remove team members, use twg whoami for the current authenticated user, and see richer credential and search-code connection diagnostics in twg doctor.
• Richer command guidance: notifications can include expanded content, while help surfaces inherited Bitbucket options, route suggestions, and clearer input errors.
• More reliable installation and Bitbucket workflows: fixes improve asynchronous merge waiting, pipeline-log diagnostics, token setup, Windows architecture selection, and installer distribution configuration.
• Clearer discovery and reporting: twg user get me identifies the current user, twg jira space query --first is available again, twg work query rejects a misleading --limit use with repair guidance, and partial work-search output identifies reduced source coverage.
• Safer validation: twg jsm knowledge-base search query rejects a missing project locally, while Confluence space instructions handle empty responses consistently and accept only Markdown or HTML body formats.

Command changes

The following product-organized tables document public CLI command changes, showing the previous and current form of each affected command. Unless noted, existing options and positional arguments still apply.

Assets

Bitbucket

--workspace and repeatable --repo continue to work before or after Bitbucket subcommands.

Confluence

Goals

Projects

Jira

Jira Service Management

Meetings

Pull requests

Search code

Subgraph

Teams

User

Rovo

Compatibility notes

• Non-interactive writes: automation, JSON output, and sessions without a controlling terminal must provide required values explicitly; destructive operations require --yes, and consent requires --agree or --yes.
• Confluence body files: twg confluence content get --body-output-file is removed; use --body-only --output-file <path>.
• JSM inline JSON: migrated JSM commands use --input-json or --variables-json; --input-file and --variables-file are no longer available where those inline inputs apply.

1.1.1 — 2026-07-24

Highlights

• Confluence authoring: add inline --edits <json> updates, improve OAuth space creation reliability, and clarify dry-run snapshot reconciliation.
• Jira administration: add top-level space queries and project lifecycle commands.
• Notifications and responsibility routing: expand grouped notifications into individual items with --expand-groups, and add the responsibility-routing skill for owner and expert discovery.
• Automation reliability: make bb pipeline tail require an explicit step for agent, structured-output, and non-terminal runs; text-mode users retain the interactive picker.
• Updates and installation: make binary updates faster and more reliable, refresh stale minimum-version policies, and preserve compatibility with older skills-install invocations.

Compatibility notes

• Confluence body formats: page, live doc, and blog post creates and updates now require --ack-body-formats.
• SDK-migrated command input: twg jira board cells query, twg jsm alert query, and twg jsm alert delete now use inline JSON input; their --input-file and --variables-file options are no longer available.

1.1.0 — 2026-07-21

Highlights

• Private Confluence content: twg confluence content create --private creates creator-only pages, blog posts, live docs, whiteboards, databases, folders, and slides until you share them.
• JSM agent availability: Added commands to list project agents and view or set agent availability.
• Cross-provider code search: twg search-code now supports GitHub- and GitLab-only organizations without requiring a Bitbucket token, and guides unlinked identities through the correct authorization flow.
• More reliable results: GitHub pull-request repository filtering, partial Jira work-query results, Confluence comment resolution, whiteboard/database reads, and skills-directory discovery now handle their edge cases correctly.

Compatibility notes

• Docs command: Bare twg docs now displays help; use twg docs query or twg docs search explicitly.
• Talent position lookup: Use twg talent position get or twg talent position query instead of bare twg talent position.

1.0.25 — 2026-07-20

Highlights

• Confluence: space instructions are generally available with markdown/HTML/ADF support, and content task commands now handle complete/reopen, validation, agent output, and errors more reliably.
• Bitbucket and Kitsune: add failed-pipeline-step reruns, harden pipeline log downloads, and expand Kitsune insight update/delete workflows.
• Setup and rollups: improve Jira default-site remediation, install PATH guidance, Confluence permission validation, org-tree people reconciliation, and PR leadership time windows.

Compatibility notes

• Space instructions now default to markdown for --body-format; pass --body-format atlas_doc_format if automation expects ADF.

1.0.24 — 2026-07-13

Highlights

• User lookups: twg user bulk-lookup now matches by email by default, with explicit --email and --legacy-username selectors alongside account ID lookup.
• Org work rollups: Added twg work-tree (alias status-rollup) to summarize a manager's reporting tree activity across Jira, PRs, goals, projects, docs, and videos in one command.
• Clearer login guidance: Fixed the "authentication required" error to point directly at twg login instead of an unrelated personal API token page.
• Windows installer reliability: Fixed the installer hanging on interactive prompts, slow or failing downloads, and garbled output on Windows.
• Pipeline log search: twg bitbucket pipeline grep now retries transient log fetch failures instead of aborting the search.
• Input validation errors: Commands now report invalid input as structured errors with a consistent exit code and agent-mode diagnostics.
• Consistent status indicators: Unified the glyphs used for success, failure, warning, and pending states across CLI output.

Compatibility notes

• Auth failure error codes changed: missing-credential failures now exit with code 77 (previously 1) and report error.code = "AUTH_REQUIRED"; expired/rejected-credential failures now report error.code = "AUTH_INVALID" (previously AUTH_EXPIRED, and now also covers revoked/wrong-scope/malformed tokens). Update any automation keyed on exit code 1 or the exact string AUTH_EXPIRED.

1.0.23 — 2026-07-10

Highlights

• Login and setup unification: Interactive prompts now consistently read from your controlling terminal — bash <(curl …) and curl | bash work identically, and CI/detached environments get clear guidance pointing at TWG_TOKEN/TWG_USER env auth. Agents see targeted no-terminal errors directing them back to a PTY.
• Bitbucket setup and installer fixes: Windows and Unix installers now behave identically for the Bitbucket token confirmation. Added bb alias for twg setup bitbucket, and fixed twg doctor hints to point at the correct setup commands instead of unrelated login paths.
• Agent plugin lightweight setup: Codex, Claude Code, and the new Cursor plugin now ship as lightweight setup bridges that install or reuse the hosted CLI and global skills instead of bundling a full runtime.
• Jira board and sprint cleanup skill: New bundled twg-jira-resolve-merged-work skill for dry-run-first cleanup of boards, sprints, epics, or spaces using merged pull request evidence.
• Improved error handling and output: Bitbucket and Confluence commands now report client-side validation failures as structured errors. --output-summary stays compact by retrying side files or emitting bounded YAML instead of dumping full payloads.
• Auth token improvements: TWG_TOKEN is now sent as raw Bearer auth, with TWG_USER also set for Basic auth.
• Notifications: twg notifications no longer shows misleading grouped counts — output now indicates only whether a row has collapsed events.
• New Trello capability: trello card create and trello card update now support card descriptions.

Compatibility notes

• The top-level --save flag has been removed; use twg setup default-site instead.
• The --no-browser flag and clipboard token paths are no longer available; use TWG_TOKEN environment variable for non-interactive auth.

1.0.22 — 2026-07-09

Highlights

• Jira workitem comments: Added visibility controls for project roles and groups, plus detailed transition failure information and field metadata in JSON output.
• Jira workitem artifacts: New twg jira workitem link and twg jira workitem unlink commands for managing remote issue links.
• JSM integration: twg jira workitem create now detects Service Management projects and guides users to twg jsm request create; new twg jsm sla get and twg jsm sla workitems commands for SLA management.
• User lookup: New twg user bulk-lookup command for batch lookups by username or account ID.
• Login timeout: Interactive login now times out after 5 minutes (2 minutes in CI/agent mode) with fast-fail guidance for non-interactive shells.
• Resolve output: twg resolve --output json now includes the canonical URL in results.

Compatibility notes

• jira workitem get and jira workitem transition now query only the targeted site and return clear single-site errors instead of scanning other accessible sites; Jira issue ARIs are treated as explicit single-site targets, and using an ARI/URL together with a conflicting -s flag is rejected.
• Removed the hidden twg confluence page get compatibility alias.

1.0.21 — 2026-07-02

Highlights

• JSM service management coverage: Add twg jsm sla metrics for listing configured SLA metrics on a project, and make twg jsm request create easier to use with friendly service desk and request type resolution plus clearer required-field guidance.
• Simpler agent skill setup: twg skills install, twg setup, twg upgrade, and twg doctor now use one canonical home skills location and the same repair command, reducing stale-skill confusion across agents.
• More reliable terminal behavior: OS credential agents such as ssh-agent and gpg-agent no longer cause false agent-mode detection, and Windows background upkeep no longer flashes a command window every few minutes.
• Sharper agent guidance: Bundled skills now include clearer guidance for Assets lookup, batched user discovery, canonical work/project/goal/Jira aliases, and compact outputs that reduce follow-up inspection turns.

Compatibility notes

• twg skills install --local is no longer supported. Use twg skills install for the canonical home install, or agent-specific options such as twg skills install --agent claude when needed.

1.0.20 — 2026-07-01

Highlights

• GitHub Rovo search results are restored: twg search and twg rovo search --app github now return GitHub-backed results again instead of an empty result set.
• Clearer Confluence HTML authoring guidance: twg confluence content body-formats html now uses the canonical @atlassian/atlassian-html-format reference, including the full text and highlight colour palettes.
• Updated agent marketplace presence: Codex and Claude Code plugin marketplace copy now uses the Teamwork Graph tagline, refreshed logo assets, the current website link, and work-context starter prompts.

Internal preview

• Add an internal-only preview of twg jira workitem similar for finding semantically similar Jira work items from an item's title and content.

1.0.19 — 2026-07-01

Highlights

• Bitbucket pipeline automation: Add twg bitbucket pipeline trigger to start branch, pull-request, commit, and custom pipeline runs. It supports safe environment-backed secured variables, duplicate-key validation, optional terminal polling with --wait, and build URL output with --open.
• Broader Confluence authoring: twg confluence content now supports whiteboards and databases. Comment creation uses the newer Confluence agentic endpoint and accepts --body-format plus --body-file, while HTML export preserves existing data-local-id attributes by default.
• Multi-provider code search: twg search-code can filter readiness and results across Bitbucket Cloud, GitHub, and GitLab with repeatable --app, includes authentication URLs in text and JSON output, and adds twg cache clear --code-search-auth-status.
• Richer Atlas and access data: Atlas goals now read and write descriptions, Atlas projects can include their "about" details with --include-description, and twg access --bitbucket discovers workspaces through the supported Bitbucket workspace endpoint.
• Better automation and setup UX: JSON and agent-mode failures no longer emit duplicate error objects, twg logout returns a structured result in JSON/agent mode, Windows success-path progress is no longer rendered as PowerShell errors, and the Codex and Claude Code plugin bundles now reuse the public TWG installer.
• SDK and agent tooling: Add the runtime-first @atlassian/twg-tools package with command definitions exposed from @atlassian/twg-tools/defs, add shorter @atlassian/twg-just-bash browser export names, and bundle the twg benchmark lite single-prompt comparison workflow.

Compatibility notes

• twg confluence space instructions get|set replaces the visible agents-context wording. The old surface remains as a hidden deprecated alias.
• twg loom settings get is hidden from external builds because the backing Loom settings field is internal-only.
• @atlassian/twg-just-bash is independently versioned and moves to 1.1.0; @atlassian/twg-tools moves to 0.2.0.

1.0.18 — 2026-06-25

Highlights

• Complete Confluence permission management: twg confluence content permissions now uses the Confluence v2 API and adds filtered, paginated list, full replace, and filtered remove workflows alongside add and clear. Destructive changes require confirmation and retain server-side self-lockout protection.
• Restriction-state controls and clearer automation: New twg confluence content restriction-state get|set commands manage OPEN, EDIT_RESTRICTED, and VIEW_RESTRICTED states. Permission envelopes now report command-specific resource types, and comment deletion explains when child replies must be removed first.
• More resilient discovery: twg responsibility infer returns partial results when traversal reaches a cloud without the relevant active product, while twg jsm portal query supplies required defaults and limits unfiltered results to three.
• Correct Talent site targeting: twg talent position query|me --site now resolves the selected site's cloud ID instead of reusing stale global context; --site and --cloud-id are mutually exclusive.

Compatibility notes

• Confluence v2 permissions identify users by account ID and groups by group ID rather than group name. The legacy v1 restriction service has been removed.

1.0.17 — 2026-06-24

Highlights

• Smoother Confluence content workflows: twg confluence content get can now write editable page content directly to a local file with --body-output-file, while keeping the snapshot token available in JSON output.
• More reliable pasted-link support: Confluence content lifecycle commands now accept full Confluence URLs, tiny links, ARIs, and numeric IDs consistently, including move targets such as --parent-id, --before, and --after.
• Clearer Confluence URLs and comment metadata: create/update output now prints full content URLs when available, and inline comment responses expose inlineMarkerRef for clients that need to reconnect comments to document positions.
• Better troubleshooting and upgrades: backend-error feedback links are easier to open from terminals, twg feedback --last-error exposes the most recent error context, Windows twg upgrade now runs in the foreground for clearer progress and failure reporting, and twg upgrade --version <version> now installs the requested version instead of printing the CLI version.

1.0.16 — 2026-06-24

Highlights

• Bounded work and pull-request reporting: twg work query now accepts --from and --to, and pull-request activity queries support upper date bounds. Help documents accepted date formats and reviewer-count semantics.
• More reliable discovery: twg work query no longer fails when --to is omitted, and twg help <terms> searches the full help index even when a term also matches a command namespace.
• Better agent output and routing metadata: Pull-request queries include counts in compact --only-counts output, while graph-backed, Rovo Search, Collaboration Graph, mixed-family, and compatibility commands expose corrected tier metadata.

1.0.15 — 2026-06-23

Highlights

• General availability: Remove beta product labels from CLI and DAC surfaces. Beta terms now apply only to beta-channel releases, and upgrade guidance consistently points to twg upgrade.
• Search that fails gracefully: User search now limits each Jira, Confluence, and Slack source to 30 seconds and returns partial results when one times out. Rovo Jira search adds actionable filter suggestions after transient search-service failures.
• Safer automation contracts: twg jira workitem create accepts --yes, Teams create and update default scope to the configured site, and local validation failures return structured 400-level usage errors with exit code 2.
• Reliable Trello identifiers: Read commands accept canonical ARIs from earlier results while preserving supported shortlinks and usernames. Malformed or wrong-entity board and member identifiers now fail locally with clearer guidance.
• Public feedback and steadier Windows setup: twg feedback opens an anonymous, prefilled public support form, backend errors link directly to it, and the Windows installer now handles successful setup commands that write diagnostic output.

Compatibility notes

• The deprecated twg confluence space query alias has been removed. Use twg confluence space list.

1.0.14 — 2026-06-23

Highlights

• Run Bitbucket pipelines from the CLI: Add twg bitbucket pipeline run for branch, tag, and custom pipeline executions.
• Complete Confluence draft workflows: Add --draft to content create, get, and update. Purge errors now explain the required trash-first flow, and singular label and comment aliases open the correct help.
• Safer Jira operations across sites: Transitions now identify issues found on another accessible Jira site and require an explicit --site. Work-item deletion accepts either a positional issue key or --id and documents its permission requirements.
• More reliable lookup and recovery: Resolve exact Jira, Atlas project, and Atlas goal keys with optional type disambiguation. Ambiguous org-tree searches return structured candidates instead of failing, and Assets rejects malformed object IDs before making requests.
• Clearer agent work summaries: Status-rollup guidance now covers personal work summaries and defines reviewed pull-request counts consistently.

Upgrade notes

• Public installs and upgrades now use the signed twg or twg.exe executable directly and clean up legacy twg-bin and twg.cmd launchers automatically. Continue upgrading with twg upgrade.

1.0.13 — 2026-06-22

Highlights

• Goals type discovery: Add twg goals types and let twg goals create resolve --goal-type <name> safely. Ambiguous names list matching IDs, pagination fails closed instead of returning partial results, and invalid selectors fail before authentication or network calls.
• Confluence drafts and browser URLs: Add publish and delete workflows for page and blog-post drafts. twg confluence content get now accepts browser URLs for folders, whiteboards, databases, and dated blog posts, while command-local --site reliably controls lifecycle operations.
• Safer Confluence permissions: twg confluence content permissions add now previews the target, operations, and grantees before applying restrictions and requires --yes for non-interactive use.
• Bitbucket authentication repair: twg setup bitbucket --force now reliably starts re-authentication. The command also supports --no-browser, including clipboard-based token setup.
• Actionable diagnostics: Fatal backend errors now provide a clickable, prefilled public feedback form link with safe command and trace context, so users can report failures without running another command or copying user-provided arguments.

Compatibility notes

• Compass commands and Loom workspace query are now hidden from public builds. twg loom get examples now correctly use positional IDs.
• Direct-install updates do not support downgrading the new one-binary layout back to pre-GA wrapper-era versions. Rollbacks should target supported GA one-binary releases.

1.0.12 — 2026-06-21

Highlights

• Bitbucket pipeline diagnostics: Add twg bitbucket pipeline tail for live step-log streaming with WebSocket notifications and polling fallback, plus pipeline grep for literal or regex matching with case, context, match-limit, and step filters. pipeline get --logs also accepts step UUIDs and --full; streaming output is available as JSON and JSONL for agents.
• Work and document discovery: Redesign twg work query as an authored-by-default relationship summary with full-window counts, bounded previews, activity selectors, and separate document and calendar sections. Add twg work search, source-balanced broad Rovo and document search, and Rovo-backed Trello board and card search.
• Jira search and issue types: Route fuzzy work-item text searches through Jira-native JQL, add explicit issue-type ID and ARI lookup through twg jira workitem types get --id, and bound issue-type discovery to project-scoped selectors.
• More reliable relationships and projects: twg context get and twg responsibility get|infer now discover direct relationships without metagraph dependency while preserving typed enrichment and reporting scan coverage. Project queries preserve inherited filters and accept repeatable --name organization roots.
• Resilient partial results: Confluence tree reads now retain accessible descendants when another descendant is inaccessible or untitled. twg pull-requests query independently hydrates author and reviewer relationships, keeps valid deduplicated pull requests when individual relationships fail, reports partial omissions, and calculates aggregate counts independently of --limit.
• Bitbucket credential repair: twg setup bitbucket --force now re-persists an unchanged token, repairing stale or partially stored credentials instead of silently leaving authentication broken.

Compatibility notes

• twg jira workitem types query --id remains available as a deprecated compatibility path; use twg jira workitem types get --id. Issue-type queries without an ID now require a project-scoped selector rather than enumerating every visible Jira issue type.

1.0.11 — 2026-06-18

Migration required

• twg code-intelligence is now twg search-code, with no compatibility alias. Update scripts and agent skills that use the old command; its subcommands, flags, and behavior are unchanged. Confluence version listing also removes --sort and --content-type because history is always newest-first and content type is resolved automatically.

Highlights

• Rovo connectors: twg rovo list-apps gains the list-connectors alias, compact readiness and remediation output, structured JSON failures, twg rovo auth <app>, and authentication preflight for explicit connector searches. Salesforce search is now available through --app salesforce or --app sfdc, including --type narrowing.
• Confluence history: twg confluence content versions|history now supports pages, live docs, and blog posts, accepts Confluence URLs, renders historical bodies as Markdown or HTML, and adds server-computed diff plus version restore workflows.
• Discovery and administration: twg context get <reference> resolves canonical Atlassian ARIs into a product-neutral context envelope. Assets adds schema-scoped object-type discovery, schema settings, UUID reference IDs, typed select options, and richer attribute metadata; JSM approval commands are now available in every build profile.
• Reporting and Bitbucket: twg talent position query --group-by produces up to three levels of human-readable headcount aggregation, and Bitbucket PR create, update, and comment commands can upload images with fail-fast placeholder validation.
• Reliability and cleanup: transient upstream failures are retried and surfaced with stable error codes and trace IDs instead of empty results. twg uninstall now removes scheduled upkeep, installer-owned skills, credentials, and local TWG configuration without leaving update checks behind.

1.0.10 — 2026-06-15

Highlights

• Safer OAuth maintenance. Adds non-interactive twg auth refresh with persistence checks and cross-process coordination so sandboxed agents cannot lose rotated refresh tokens. New twg upkeep schedules near-expiry refreshes and update reminders; setup enables it automatically and uninstall removes it.
• Richer Confluence content workflows. Adds Word export alongside PDF, routes content listing through the unified endpoint, and improves SVG and CSV handling for whiteboards and databases.
• Better Bitbucket CI diagnostics. Adds pipeline test report and test case retrieval, while twg doctor now verifies live Bitbucket connectivity instead of only checking whether a token exists.
• Faster agent discovery and more tolerant reads. twg help can discover bundled skill references and skill: targets, Jira and Confluence guidance is split into focused skills, and common PR, Assets, media, user, meeting, and JSM reads accept more natural command shapes with structured fallback guidance.
• Broader scoped access and steadier output. Expands OAuth scope mappings for Confluence templates, drafts, properties, and Jira administration, while improving row output, cloud ID resolution, and degraded GraphStore handling.

Upgrade Notes

• twg confluence content export <id> now requires --format word or --format pdf; the bare command no longer defaults to PDF.
• twg setup now attempts to enable scheduled upkeep. If the OS scheduler is unavailable, setup still completes and prints twg upkeep enable as the repair command.

1.0.9 — 2026-06-12

Highlights

• Adds twg compass for operational component data and twg responsibility (twg ownership) for ranked ownership, expertise, review, and escalation discovery.
• Expands product workflows with JSM approvals, Teams membership management, Jira administration, Assets reference types, Confluence copy and templates, Bitbucket file reads, and file-backed pull request descriptions.
• Makes Rovo connected-app search more reliable across Graph and federated connectors, including Slack, surfaces downstream failures instead of misleading empty results, and adds connector auth state, targeted recovery, and twg rovo auth <app>.
• Adds twg uninstall with safe preservation and explicit purge modes, and retires the beta custom Homebrew distribution path.
• Improves embedded and agent operation with delegated OAuth via TWG_OAUTH_ACCESS_TOKEN, fresher saved auth profiles, TWG_SKILLS_DIR, twg --report-paths, command-tier metadata, and Code Search enablement checks.
• Hardens Jira and Confluence writes with REST-backed Jira updates, ADF mentions, reliable inline comment defaults, and warnings for local image references that cannot be uploaded automatically.

Compatibility Notes

• Removes deprecated JSM request-location and priority-group commands, plus the legacy twg confluence whiteboard and twg confluence database groups. Use unified twg confluence content ... --content-type whiteboard|database; availability remains build-gated.

1.0.8 — 2026-06-08

Patch Changes

• Fix Windows sign step: use npm instead of pnpm to bootstrap tsx (pnpm not on PATH on Windows runner).

1.0.7 — 2026-06-08

Patch Changes

• Pipeline: each step now self-installs via pnpm install --frozen-lockfile — fixes EMFILE on macOS runners after pnpm migration.

1.0.6 — 2026-06-07

Patch Changes

• Updated pnpm to 11.5.2 and ensured corepack enable pnpm is present in all release pipeline steps.

1.0.5 — 2026-06-06

Highlights

• Confluence legacy command deprecation. The twg confluence whiteboard and twg confluence database command groups are deprecated. Use twg confluence content create|update|get --content-type=whiteboard|database instead.
• twg logout now fully revokes tokens. Signing out revokes both the OAuth access and refresh tokens server-side, not just clears local credentials.
• New command: twg confluence space me. Resolves the authenticated user's personal Confluence space directly.

Breaking Changes

• Confluence body format plain removed. Page, blog, and content write commands no longer accept plain or text as a body format. Use markdown, md, or html.
• Jira workitem create now uses ADF for --description. The twg jira workitem create --description flag now accepts Markdown or HTML and converts to Atlassian Document Format, consistent with workitem update.

1.0.4 — 2026-05-30

Patch Changes

• Version-only release prepared for 1.0.4 without pending changesets.

1.0.3 — 2026-05-30

Patch Changes

• 5f21300: twg login now uses OAuth by default in installer paths too. The public curl installer (bin/install-template.sh) and Windows installer (bin/install-template.ps1) no longer forward --from-clipboard to the post-install login step. PAT remains available as a hidden dev escape hatch via twg login --pat. --force always re-authenticates with the default flow (OAuth) unless --pat is also passed. Non-OAuth-capable contexts (non-TTY, the TWG_TOKEN+TWG_USER env shortcut, twg env auth, and existing PAT auth.conf users not passing --force) continue to use token flow.
• 3ceaa91: Remove beta labels from version output, top-level help, DAC navigation and pages, and consent copy for GA readiness.
• c31202b: Do not ask for email or site during twg setup OAuth login; keep the setup profile prompt only for the hidden PAT setup flow.

1.0.2 — 2026-05-29

Highlights

• OAuth-first login and clearer auth recovery. twg login now uses OAuth by default, --force consistently starts a fresh auth flow, --no-browser works across login prompts, and 401/403 failures now show mode-aware remediation with a stable AUTH_EXPIRED agent envelope and exit code 77. Bitbucket token setup remains token-based and is prompted when needed.
• New and expanded Atlassian work surfaces. This release adds Assets write lifecycle commands with dry-run/delete guards, Confluence content update/body-format guidance/move/label improvements, JSM customer request creation and request-type field metadata, Loom video comments, Bitbucket deployment variable mutations, repository/branch/pipeline/PR parity commands, and external Code Intelligence access.
• Better query, context, and agent behavior. Generic query contracts now distinguish natural language from query-language inputs, agent summaries and compact/evidence output are more predictable, TWG_AGENT_DEFAULTS=1 lets installed skills call twg directly with agent-friendly defaults, and twg doctor now highlights duplicate twg binaries on PATH.
• Safer defaults and richer object handling. pr-tree is now count-only by default, supports --limited-fetch, and bounds zero-flag merged rollups to the last 30 days. Jira issue-key resolution, Jira rich-text validation, Talent position lookup/rendering, JSM linked repositories, and Teams ARI lookups were also improved.
• Release and packaging updates. Codex plugin packaging is now available with public-safe bundled guidance, the benchmark latest-release pipeline can run non-interactively, and release builds hide or remove unfinished/unsupported Jira Align, Trello, Talent proposal, and JSM channel/request-location commands.

Upgrade Notes

• twg login is OAuth-first. The old --oauth flag has been removed because OAuth is now the default path. Existing classic token auth still works for installer, clipboard, non-TTY, environment-token, and existing auth.conf flows.
• Stale OAuth environment overrides (TWG_OAUTH_CLIENT_ID, TWG_OAUTH_AUTH_BASE_URL, TWG_OAUTH_SCOPE) were removed; endpoints and client ID now come from environment config.
• Some in-development or unsupported commands are intentionally absent from release builds, including Jira Align, JSM channel mutations, JSM request-location query, Trello bulk list deletion, and Talent focus-area proposal commands.

1.0.1 — 2026-05-21

Highlights

• Adds twg api to every build profile, so signed-in users can make authenticated Atlassian REST and GraphQL calls with the CLI's saved credentials, product shortcuts, and structured JSON/JSONL output.
• Makes twg admin available in the external CLI with Admin API key onboarding, org switching, user lifecycle commands, and lifecycle capability checks.
• Improves command discoverability across Jira, Confluence, JSM, Bitbucket, Trello, Loom, Talent, Atlas Goals, and Atlas Projects with clearer help, examples, guards, synonyms, and safety guidance.
• Adds twg confluence ... --help-body-format so rich-body commands can show supported plain text, Markdown, and HTML input formats directly from the CLI.
• Hardens day-to-day workflows, including JSM service search, Confluence tiny URL reads, Jira goal/project links, Ctrl+C cancellation during long fetches, OAuth refresh timing, and final analytics delivery before exit.

1.0.0 — 2026-05-19

TWG CLI 1.0.0 brings the 0.9.x release line together as a stable baseline for agent-oriented Atlassian workflows, with broader Jira and Confluence coverage, more reliable installs and updates, improved authentication, and better command discoverability.

Highlights

• Establishes the agent-friendly command model as the stable 1.0 baseline, with richer Jira, Confluence, Bitbucket, Loom, and Feedback workflows.
• Adds a large Confluence content surface that lets agents work with pages, blog posts, whiteboards, databases, folders, comments, labels, permissions, tasks, versions, history, attachments, and reactions through a more uniform command shape.
• Expands Jira REST parity across work items, attachments, worklogs, comments, properties, changelog, votes, filters, dashboards, boards, sprints, project components, versions, and bulk operations.
• Improves install, update, and OAuth behaviour so the binary is safer to distribute and easier for agents to diagnose.

Breaking changes

• No intentional command removals are included in this release. Some Confluence and install flows now prefer normalized command shapes and clearer guarded behaviours, while compatibility aliases remain where required.

Confluence

• Adds the new twg confluence content surface as the preferred agent-friendly entry point for content lifecycle operations.
• Wires page, blog post, whiteboard, database, and folder lifecycle parity where underlying services exist, including page create/update/archive/restore/delete and move, blog post create/update/delete, whiteboard/database list/get/create/delete, and folder get/create/delete.
• Adds content subresources for:
  • attachments upload/download.
  • comments list/get/create/reply/update/resolve/reopen/delete.
  • labels list/add/remove.
  • permissions list/add/remove.
  • tasks list/get/complete/reopen.
  • versions and history list/get/restore.
  • reactions get/add/remove.
• Makes --type optional for many ID-only confluence content commands by adding dispatcher probes and structured unsupported-operation errors.
• Adds confluence content create over the agentic content API for pages, live docs, blog posts, whiteboards, databases, embeds, smart links, folders, and slides.
• Normalizes twg confluence space verbs to align with the new content surface, while preserving compatibility aliases where needed.

Jira

• Adds REST-backed Jira work item subresource commands for attachments, comments, watchers, issue links, properties, changelog, votes, and worklogs.
• Adds twg jira workitem bulk-get and twg jira workitem bulk-transition, including a safe dry-run default for bulk transitions.
• Speeds up repeated twg jira workitem get reads by using REST search batching when possible and falling back to per-issue reads when needed.
• Adds rich description/comment/worklog input formats for plain text, Markdown, HTML, and ADF, with HTML as the safer default for rich-text authoring.
• Adds REST-backed Jira filter and dashboard commands, including saved filter lifecycle, sharing/subscriptions, dashboard lifecycle, and gadget administration.
• Adds Jira Agile board and sprint parity commands, sprint snapshots, board and sprint discovery from project URLs or issue keys, and project component/version lifecycle commands under jira space.

Bitbucket, Loom, and Feedback

• Adds twg bb inbox for a bounded current-user PR inbox with reviewer and author filters plus compact task, comment, and check signals.
• Adds twg bb pipeline latest-failure to locate the newest failed pipeline and hydrate failed step log tails in one bounded flow.
• Normalizes Bitbucket PR descriptions that contain escaped newlines so agent-generated Markdown renders correctly.
• Adds Loom transcript preview and file-backed full transcript support to twg loom get.
• Routes twg feedback Jira issue creation through the Jira REST gateway with clearer authentication and timeout guidance.

Agent and workflow experience

• Adds agent-friendly command repairs and aliases for common lookups, including positional user search, pull-request status aliases, project/goal shortcuts, Confluence page positional get, and Loom/video/meeting URL hydration.
• Splits bundled TWG workflow guidance into outcome-specific skills for better agent discovery.
• Restores and refreshes generated CLI catalog discoverability for the new Confluence content and space surfaces.

Auth, install, and update hardening

• Ships and fixes the OAuth device-code login path, including refresh behaviour and a corrected scope allowlist for the TWG CLI OAuth client.
• Adds low-frequency update checks from doctor/version commands and reports installed TWG skill freshness with repair guidance.
• Fixes Windows twg upgrade replacement flows, records install metadata, and preserves logs when background replacement fails.
• Warns about duplicate Homebrew/direct installs and prevents accidental direct installs over active Homebrew-managed installs.
• Simplifies direct installers by requiring setup finalization and sharing the POSIX installer template across shell installs.
• Fixes shell installer checksum parsing for CRLF SHA256SUMS manifests.
• Improves command compatibility coverage so supported aliases resolve more reliably.

0.9.8 — 2026-05-14

Minor Changes

• b48bdbd: Add twg access for product, site, org, test-like, configured-site, Bitbucket, and Rovo access inventory checks.

• 91c25e2: Add a BUILD_ENABLE_CONFLUENCE_CONTENT_TOOLS build flag so release-style setups can opt into the agentic Confluence content tools bundle. The bundle currently exposes a confluence content get preview backed by Confluence's /wiki/api/v2/agentic/content/{id} (and ?content_url=) endpoint, with summary, outline, and full detail levels; future agentic create/update/delete commands will land behind the same flag so the surface graduates as a unit.

  The flag defaults to on in dev/internal builds and off in release/external builds; invalid values warn and fall back to the build-profile default. Set BUILD_ENABLE_CONFLUENCE_CONTENT_TOOLS=1 at build time (e.g. BUILD_ENABLE_CONFLUENCE_CONTENT_TOOLS=1 npm run setup) to opt a release-style build into the surface.

  Coordinate the default flip with the monolith's agentic_content_tools Statsig gate so client-side availability tracks server-side availability.

• 2f94fe9: Help discovery: typed Guards: and Synonyms: sections, plus ranking fixes

  Adds a Guards: heading and an inline Synonyms: line that authors can drop into any command's addHelpText("after", …) block. The catalog parses both into typed HelpCommandRecord.guards and HelpCommandRecord.synonyms fields, the index persists them, and flexsearch slots them into the existing weight ladder (path: 12, cmd: 10, synonyms: 9, alias: 8, meta: 5, desc: 3, guards: 2, args: 2, opts: 2, input: 2):

  • synonyms — weight 9, just above alias: 8, so curated semantic neighbours such as cmdb → assets or calendar → meetings win against incidental description prose and even against legacy aliases when the alias is the wrong target.
  • guards — weight 2, low enough to act as a tie-breaker for sibling-overlap commands without inventing matches on unrelated commands.

  As part of the content rollout, pr-tree and workitem-tree had their existing free-form prose footers converted to typed Guards: blocks. Those footers were previously not indexed at all (the catalog only captured Examples: blocks), so the conversion is also a strict gain in search coverage for those two commands.

  Two complementary ranking fixes ship in the same change so existing commands benefit even without authored content:

  • Drop single-character query tokens (a, i, o, …) at the tokenize boundary so noise tokens stop prefix-matching every assets * / auth * path.
  • Damp the path-prefix bonus by the ratio of query-token length to path-token length so short prefixes no longer dominate ranking against full-word matches.

  guards and synonyms are additive optional fields, so the help-record and help-index schema versions stay at 1. Cache invalidation on upgrade is carried by programSignature(), which now hashes per-command rendered help-text length and so changes whenever a Guards/Synonyms block is added or removed.

  Authored content covers the full multi-surface ambiguity surface: meetings, docs, videos, spaces, org-tree, pr-tree (existing prose footer converted), workitem-tree (converted), assets, collaborators, recently-viewed, notifications, pull-requests, context jira workitem, context user, context confluence (page/space/blogpost/whiteboard), goals, projects, focus-areas, talent, teams, jsm incident query, jsm alert query, jsm post-incident-review query, confluence blog, confluence database, confluence whiteboard, bitbucket search prs, and bitbucket default-reviewer.

  Examples of fixed rankings (twg help <freeform> → top result):

  • find a persons meetings → meetings (was assets subtree)
  • okr → goals
  • postmortem → jsm post-incident-review query
  • outage incident → jsm incident query
  • monitoring page alert → jsm alert query
  • issue context → context jira workitem
  • user perimeter → context user
  • team membership → teams
  • open positions → talent
  • blog post announcement → confluence blog
  • whiteboard diagram → confluence whiteboard
  • default reviewer → bitbucket pull-requests default-reviewer
  • find pr by text → bitbucket search prs
  • atlas project → projects
  • focus area initiative → focus-areas
  • cmdb → assets
  • who reports to whom → org-tree

• 15a40ce: Make twg work query count-first by default, add distinct entity counts for validated sections, require --fetch items when callers need hydrated work item rows, and guard unbounded item fetches with a 2000-relationship preflight limit.

Patch Changes

• 3b0a2d7: Add a --body-output-file option to Confluence page reads so body content can be saved to a file while JSON output stays metadata-focused.
• e3c4f17: Stop user search from calling the admin users GraphQL lookup.
• ab03591: Document agent-facing help search behavior changes that de-rank advanced raw graph/debug commands unless the query names them directly, and tighten high-fanout agent output contracts for org, goal, and project hydration commands.
• 7f69e9e: Add project-aware Jira field discovery and generic custom field support for workitem create and update.
• ba5649a: Stop publishing the duplicate Windows setup.ps1 installer; install.ps1 is now the canonical PowerShell installer URL for both public Bifrost and internal Statlas releases.
• 7f2d4f9: Skip the optional Bitbucket token prompt during login when accessible product discovery confirms the account has no Bitbucket access.
• b437a97: Add twg workitem-tree with an issue-tree alias to summarize Jira work item activity by reporting tree with fast GraphStore Cypher-backed counts by default and guarded --full-fetch issue sample fetching.

0.9.7 — 2026-05-06

Patch Changes

• 5087758: context: gracefully handle partial GraphQL responses; search snippet formatting

  twg context user, twg context confluence <type>, and twg context jira workitem now return whatever data the graph store did manage to fetch when one or more relationship branches fail to load. The dropped branches are surfaced as a brief warning on the v2 envelope's warnings field (and to stderr in text mode) instead of throwing a hard Unable to retrieve data: cannot route the request error.

  twg search results no longer show multi-line raw content in snippet previews — descriptions are now collapsed to a single readable line and dimmed for visual consistency with other result metadata.

• 7f0e519: test(coverage): use real fixtures for confluence labels attachment and confluence labels custom-content

  The static read coverage harness was passing a Confluence page ID to confluence labels attachment and a placeholder 0 to confluence labels custom-content. Both produced HTTP 404s, which the harness mis-classified as "token lacks Confluence scope" — masking the fact that both commands work correctly. The harness now discovers a real attachment and custom-content ID via CQL on each run, and the Confluence skip-matchers are reordered so genuine 404s can no longer be reported as scope failures.

• 142758c: fix(search): recover full results when AGG returns partial success with downstream errors

  twg search query --assignee <user> (and other filtered searches) previously showed a warning that results were limited to Confluence and Jira. This was caused by two issues:

  1. Partial AGG responses were discarded: When AGG fanout to a downstream connector returned a 403, AGG still returned HTTP 200 with valid search results alongside a DOWNSTREAM_ERROR GraphQL error. graphql-request's default error policy treated this as a failure and threw away the usable data, triggering the Confluence/Jira-only fallback. Fixed by catching the ClientError directly in runAggSearch and recovering the data when response.data.search.results.edges is non-null, scoped strictly to the Rovo search path so other GraphQL callers are unaffected.
  2. Default search fanned out to all connected apps implicitly: Searches without --app were fetching entity types from all 55+ connected third-party apps and including them in the search, causing AGG to fan out to connectors that may fail. Third-party entities are now only included when explicitly requested via --app.

• 5f2273e: Remove jsm conversation by-container query, jsm conversation workspace experience-config get, jsm conversation workspace related-experience-config get, and jsm automation task-agent query from the public CLI surface at the request of the upstream jsmChannels owners. Also hide jsm incident get from --help, the generated catalog, and the command-policy schema until the required read:incident:jira-service-management scope is available in the twg login granular-scope flow; the runner remains wired so existing scripts continue to work.

• e189268: Add twg pr-tree with a pull-requests-tree alias to summarize pull request activity by reporting tree with fast GraphStore Cypher-backed counts, guarded full-fetch samples, and per-person direct examples.

Earlier Releases

This public changelog starts at 0.9.7. Earlier internal and pre-stable entries are omitted to keep release notes focused on currently relevant TWG CLI behavior.

Please use the latest stable installer for supported setup and update flows.