// Use when the user wants to fix a Sentry issue, auto-repair a bug from Sentry, or create a fix PR for a Sentry error. Triggers on "修复 sentry", "fix sentry issue", "sentry 修复", "sentry fix".
| name | sentry-fix |
| description | Use when the user wants to fix a Sentry issue, auto-repair a bug from Sentry, or create a fix PR for a Sentry error. Triggers on "修复 sentry", "fix sentry issue", "sentry 修复", "sentry fix". |
Fix a specific Sentry issue: gather context, confirm with user, implement fix, write regression test, verify, and create PR.
<ISSUE-ID> (required): Sentry issue short ID, e.g., TEAMCLAW-3 or TEAMCLAW-REACT-2GThe issue ID is passed as the skill argument. Example: /sentry-fix TEAMCLAW-3
| Project | Sentry Slug | Platform | Test Command | Lint Command |
|---|---|---|---|---|
| Rust backend | ucar-inc/teamclaw | Rust | cargo test --manifest-path src-tauri/Cargo.toml | cargo clippy --manifest-path src-tauri/Cargo.toml -- -D warnings |
| React frontend | ucar-inc/teamclaw-react | React | pnpm test:unit | pnpm typecheck && pnpm lint |
Determine which project based on the issue ID prefix:
TEAMCLAW-<suffix> (no "REACT") → Rust backendTEAMCLAW-REACT-<suffix> → React frontendsentry issue view <ISSUE-ID> --json
From the stack trace and error message, identify the relevant source files and functions in the codebase.
Read those source files to understand the code context around the crash/error site.
Perform local root cause analysis based on the stack trace + source code. Produce a clear root cause summary.
If the stack trace is insufficient to identify the root cause: Stop and ask the user for additional context about the issue before proceeding. Do NOT guess.
Present a fix plan to the user in the terminal. The plan MUST include:
STOP HERE and wait for user confirmation. Do NOT proceed until the user says yes.
git checkout -b sentry-fix/<issue-id-lowercase>
Example: sentry-fix/teamclaw-3 or sentry-fix/teamclaw-react-2g
Implement the fix:
Write a regression test:
#[test] in the relevant module's test section, or create a test file in the same directory. The test must reproduce the scenario that caused the original error..test.ts or .test.tsx file (create one if none exists, following existing test file patterns in the codebase). The test must verify the fix prevents the original error.Run the verification suite for the relevant platform:
Rust backend:
cargo clippy --manifest-path src-tauri/Cargo.toml -- -D warnings
React frontend:
pnpm test:unit && pnpm typecheck && pnpm lint
If verification fails:
git add <changed-files>
git commit -m "fix(<scope>): <description> (Sentry <ISSUE-ID>)"
Where <scope> is the module/component name (e.g., rag, p2p, settings, editor).
git push -u origin sentry-fix/<issue-id-lowercase>
gh pr create --title "fix(<scope>): <short description>" --body "$(cat <<'PREOF'
## Sentry Issue
Fixes [<ISSUE-ID>](<sentry-issue-permalink>)
## Root Cause
<root cause summary from local analysis>
## Fix
<description of code changes>
## Test Coverage
<description of regression test added>
## Verification
- [ ] cargo clippy / pnpm lint passes
- [ ] Tests pass including new regression test
PREOF
)"
sentry CLI is not authenticated, prompt the user to run sentry auth logingh CLI is not authenticated, prompt the user to run gh auth loginUse when the user wants to check Sentry issues, run a Sentry daily report, or monitor error trends. Triggers on "sentry 监控", "sentry 日报", "查看 sentry", "sentry report", "sentry monitor".
Create a workspace role that follows the ROLE.md specification. Use when the user wants to create a new role, refine a role definition, or turn repeated responsibilities into a reusable routed role.
Control Windows desktop applications: open/operate apps, click controls, type text, scroll, keyboard shortcuts, window management, toggle options, lists and combos, etc. Note: NEVER invoke this skill for opening web pages/visiting websites — use browser-related tools instead. Trigger words: open app, operate app, control computer, click, type, scroll, switch, dropdown, select, modify, change to, set, check, uncheck, app-control, toggle.
List all AI API keys for the current team. Shows key aliases, masked key values, and spend per key. Use when a team member asks about their API key or wants to see who has access.
Manage AI API keys and budget for the current team. Owner-only operations include adding/removing member keys and setting budget limits. Use when the team owner wants to manage AI access.
Check AI usage and spend for the current team or individual member. Shows total spend, per-member breakdown, or personal usage. Use when someone asks about costs, budget, or how much has been used.