
Aicommit
- 1 installs
- Updated July 29, 2026
- berend/my-skills
aicommit is an agent skill that creates git commits with AI-Assisted trailers and optional AI-Agent and AI-Model metadata.
About
aicommit is an agent skill that turns “commit this” into a repeatable git ritual: inspect status and diffs, stage only what belongs to the request, mirror recent oneline commit tone, and land trailers that mark the commit as AI-assisted for downstream analytics. Developers shipping with Claude Code, Cursor, or Codex benefit because attribution stays in standard git metadata instead of ad-hoc comments. Use it whenever the user asks to commit, create a git commit, or mark work as AI-assisted—typically many times per day during Build and again before Ship when you want a clean, attributable history. The skill prioritizes normal, high-quality commits over novelty messages and documents Harness Identity Resolution for AI-Agent and AI-Model trailers when resolvable. It is a workflow skill, not a hosting integration; you need a local repo and git available in the agent sandbox.
- Seven-step workflow: status, diff, staged diff, recent log, selective stage, message, commit with trailers
- Requires AI-Assisted: true trailer; adds AI-Agent and AI-Model when Harness Identity Resolution succeeds
- Uses git commit --trailer when available; falls back to manual trailer lines at message end
- Explicit rule not to stage unrelated user changes
- Matches existing repository commit message style from git log --oneline -10
Aicommit by the numbers
- 1 all-time installs (skills.sh)
- Ranked #526 of 733 Git & Pull Requests skills by installs in the Skillselion catalog
- Data as of Jul 30, 2026 (Skillselion catalog sync)
npx skills add https://github.com/berend/my-skills --skill aicommitAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 1 |
|---|---|
| Last updated | July 29, 2026 |
| Repository | berend/my-skills ↗ |
What it does
Create git commits that match repo style and append AI-Assisted trailers for later attribution and analysis.
Who is it for?
Best when you commit from the agent daily and want analyzable AI attribution without a separate commit bot.
Skip if: Repos where policy forbids AI trailers, or users who only need conventional commits with no metadata—use plain git instructions instead.
When should I use this skill?
User asks to commit, create a git commit, or mark a commit as AI-assisted.
What you get
You get a scoped commit matching repo style with AI-Assisted (and resolved AI-Agent/AI-Model) trailers ready for push or PR follow-up.
- Git commit with message matching repo style
- AI-Assisted trailer plus optional AI-Agent and AI-Model trailers
By the numbers
- Seven-step required workflow before committing
- Inspects recent commit style with git log --oneline -10
Files
AI-Assisted Commit
Use this skill when creating a git commit that should be marked as AI-assisted.
Goal
Create normal, high-quality git commits while adding stable commit trailers that make AI-assisted commits easy to analyze later.
Required Workflow
Before committing:
1. Inspect the working tree with git status. 2. Inspect the changes with git diff and, if relevant, staged changes with git diff --staged. 3. Inspect recent commit style with git log --oneline -10. 4. Stage only the files that belong to the requested change. 5. Do not stage unrelated user changes. 6. Write a concise commit message matching the repository's existing style. 7. Add the AI trailers with git commit --trailer when available. If --trailer is unavailable, place the same trailer lines manually at the very end of the commit message.
Trailer Format
Always include:
AI-Assisted: trueAlso include AI-Agent and AI-Model when they can be resolved by the Harness Identity Resolution rules.
Prefer this form when all values are available:
git commit -m "Commit message" \
--trailer "AI-Assisted=true" \
--trailer "AI-Agent=<resolved-agent>" \
--trailer "AI-Model=<resolved-model>"Omit only the unresolved trailer if either AI-Agent or AI-Model cannot be resolved.
Harness Identity Resolution
Before committing, derive trailer values from explicit coding harness context.
Use AI-Agent when the harness explicitly names the coding agent or product identity. Normalize to a stable lowercase identifier when the identity is unambiguous.
Examples:
- "You are Codex" ->
AI-Agent: codex - "You are opencode" ->
AI-Agent: opencode
Use AI-Model when the harness explicitly names the model identity. Prefer canonical provider/model IDs when the provider is explicit or obvious from the harness.
Examples:
- "based on GPT-5" in an OpenAI/Codex harness ->
AI-Model: openai/gpt-5 - "openai/gpt-5.5" ->
AI-Model: openai/gpt-5.5 - "claude-sonnet-4" in an Anthropic harness ->
AI-Model: anthropic/claude-sonnet-4
If either value is not explicitly available from the harness context, omit only that trailer. Do not infer values from general knowledge, marketing names, or user preference unless the user explicitly provides the trailer value for this commit.
Commit Message Example
Commit message
AI-Assisted: true
AI-Agent: <resolved-agent>
AI-Model: <resolved-model>Rules
- Keep the trailers at the very end of the commit message.
- Prefer
git commit --trailerover writing trailer lines manually. - Use exactly
AI-Assisted,AI-Agent, andAI-Modelas trailer keys. - Do not add
Co-authored-byunless the user explicitly asks for it. - Do not commit secrets, generated junk, or unrelated files.
- If tests or checks were run, mention the result to the user after committing, not in the trailers.
Related skills
How it compares
Use instead of improvised one-line agent commits that omit trailers and mix unrelated staged files.
FAQ
Who is aicommit for?
Developers using AI coding agents who still commit from the terminal and want stable AI-Assisted trailers for later measurement.
When should I use aicommit?
During Build while landing feature slices, during Ship before opening or merging PRs, and during Operate when iterating fixes—whenever the user says commit, create a git commit, or mark a commit AI-assisted.
Is aicommit safe to install?
It instructs git write and staging only for requested changes; review the Security Audits panel on this page and your org’s policy on AI commit metadata before use.