菜单
Run the full test suite (all packages including e2e) via `make test`. Use after making changes to confirm nothing is broken.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Operates the git-agent CLI — atomic AI commits plus co-change relations for agents, all-language, offline, no API key. Use it whenever the user wants to commit or set up git-agent; when you are about to modify a feature and need the files that historically move with it, with the commits that explain the coupling (related); when deciding which tests to run after a change (related --tests); or when checking the co-change index health (status). All related and status queries are read-only and offline (no LLM, no API key); only commit and init --scope need a provider.
Build git-agent with embedded credentials (CLIENT_TOKEN, WORKER_URL, MODEL) via scripts/build.sh. Use when you need a fully functional binary for manual testing or pre-release verification.
| name | verify |
| description | Run the full test suite (all packages including e2e) via `make test`. Use after making changes to confirm nothing is broken. |
Run the full test suite:
make test
This runs go test -count=1 across ./application/..., ./domain/..., ./infrastructure/..., ./cmd/..., and ./e2e/....
The -count=1 flag disables caching so results are always fresh. The e2e package builds the git-agent binary once at the start — any code changes made before running this skill will be compiled in automatically.
If a specific package is failing, re-run just that package:
go test ./application/...
go test ./infrastructure/config/...
go test ./e2e/...
To run a single test by name:
go test ./application/... -run TestCommitService_NoStagedChanges
If an e2e test fails, read the failing test file in e2e/ to understand the setup and assertions before investigating the source code.
Report back with the test output, highlighting any failures and their likely cause.