do-commit

name	do-commit
source	botcore
description	Run quality gates, clean up temp files, update docs, and commit changes. Two modes: quick (automated checks + commit) and full (checks + cleanup + docs gate + self-review + commit). Delegates to write-commits for message generation. Use when ready to finalize and commit changes.
version	1.0.0
triggers	["do commit","finalize","commit changes","pre-commit","cleanup","ready to commit","finalize changes"]
argument-hint	quick | full
portable	true
user-invocable	true

Do Commit

Finalize changes for commit — quality gates, cleanup, docs, and commit message generation.

Modes

Mode	What runs	When to use
quick (default)	Quality gate → commit	Small changes, confident in state
full	Quality gate → cleanup → docs gate → self-review → commit	Feature work, before PR

Quick Mode

1. Quality gate — run scripts/quality_gate.py from this skill directory
   • Detects project language(s) and runs: lint, test, typecheck, build, check-size, check-paths, circular-imports
   • Review the JSON report — if any checks fail, fix them before proceeding
2. Commit — delegate to the write-commits skill for message generation, then git add + git commit

Full Mode

Step 1: Quality Gate

Run scripts/quality_gate.py and review the JSON output. Fix any failures before continuing.

Step 2: Temp File Cleanup

Scan the repo root and working directories for artifacts that shouldn't be committed:

Pattern	What it is
*.plan.md, *.review.md	Agent work artifacts at repo root
nul	Windows null device artifact
*.tmp, *.log	Temporary/debug files
*.pyc, __pycache__/ in source dirs	Python bytecode (should be .gitignored)
Stray coverage/ dirs	Test coverage output (if not .gitignored)

Important: Confirm with the user before deleting anything unexpected. Only remove files that are clearly artifacts, not in-progress work.

Step 3: Documentation Gate

Run staleness detection first, then evaluate each:

1. Run docs_check_changelog — if stale, update CHANGELOG.md to reflect the changes
2. Run docs_check_agents — if structural changes detected, update AGENTS.md
3. Check README.md — any new commands, features, or setup steps that need documenting?
4. If skill files were modified, run skill_lint to validate

These are judgment calls — use the staleness detection results to guide whether updates are needed, but evaluate the actual content.

Step 4: Self-Review

Review the full diff (git diff --cached or git diff) for common issues:

Check	Look for
Leftover markers	Unresolved fix-me, hack, or warning markers
Debug statements	console.log, print(, dbg!, debugger
Commented-out code	Blocks of commented code that should be removed
Secrets	Hardcoded API keys, tokens, passwords
Unintended files	Generated files, editor configs, OS artifacts

If issues are found, fix them before proceeding.

Step 5: Commit

Delegate to the write-commits skill for commit message generation. The skill handles:

• Conventional Commits format
• Scope selection based on changed files
• Breaking change detection
• Multi-package commit handling

Then run git add (stage relevant files) and git commit.

Quality Gate Script

The scripts/quality_gate.py script automates deterministic checks. Run it from the repo root:

python {skill-dir}/scripts/quality_gate.py

It reads botcore.toml for per-repo thresholds and detects languages from config files. Output is JSON:

{
  "passed": true,
  "checks": [
    {"name": "lint", "passed": true, "duration_ms": 1200},
    {"name": "test", "passed": true, "duration_ms": 3400}
  ],
  "summary": "6/6 checks passed."
}

On failure, recovery suggestions are printed to stderr.

Full Checklist

See checklist.md for the expanded reference covering all checks in detail.