mitodl/agent-kit/validate-before-commit
Claude Code and agent skill by mitodl/agent-kit
- community
- ext:[{"type":"hook","description":"Pre/post lifecycle hooks"}]
Checking that the SKILL.md file is still in the repository.
Run the full validation sequence before declaring any task done. Use this skill proactively after any code or infrastructure change — run the repository's hooks with prek, then mypy and pulumi preview (where applicable), without waiting to be asked. Never declare success without passing checks.
Install
npx vskill@latest install mitodl/agent-kit/validate-before-commitSemi-extensibleMentions customization, but needs skill-specific knowledge. Read the extensible skills standard
SKILL.md
Validate Before Commit
Do not declare a task done until the full validation sequence has passed. Run these checks proactively — do not wait for the human to ask.
Standard sequence
mitodl repositories run their hooks with prek, which reads the same .pre-commit-config.yaml (or prek.toml) that pre-commit did. Use the command for the repository's prek pin:
| Repository | Hook command | |
|---|---|---|
Python with uv (prek in the prek dependency group) | uv run prek run --all-files | |
Node (@j178/prek in devDependencies) | npx prek run --all-files | |
| Neither (prek installed as a tool) | prek run --all-files | |
Not yet migrated (still uses pre-commit) | uv run pre-commit run --all-files in a uv project, otherwise pre-commit run --all-files |
In a repository with .github/workflows/autofix.yml, the same hooks run in CI as the prek check. When the hooks' only problems are ones they fixed themselves, autofix.ci commits those fixes to the PR, except under .github/. If any hook still fails, nothing is committed and the check stays red. Running the hooks locally first avoids both.
Rules
- Run the hooks before mypy; they may auto-fix formatting that would
otherwise produce mypy noise.
- If the hooks auto-fix files, stage the changes and re-run them to confirm
every hook passes cleanly.
- mypy errors are blocking — do not leave type errors for the human to clean up.
pulumi previewoutput must be reviewed: unexpected replacements or deletions
are bugs, not acceptable side effects.
For Dagster / dg changes
After modifying assets, sensors, or code location structure, also confirm the workspace loads cleanly:
Interpreting failures
Read the full error output before proposing a fix. Do not apply a speculative patch without understanding the root cause — iterating blindly on the same error multiple times is worse than pausing to diagnose.
Security
- Security scanPassed, 98 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/workflow/validate-before-commit/SKILL.md
Popularity
- Repository stars
- 5
- 7-day trend score
- 14
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Extension points
hookPre/post lifecycle hooks
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/mitodl/agent-kit/validate-before-commit)