Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/xdg/xdg-claude/how-to-commitnpx skills add xdg/xdg-claude --skill how-to-commitgit clone --depth 1 https://github.com/xdg/xdg-claudeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/xdg/xdg-claude/how-to-commit)<a href="https://agentmods.dev/skills/xdg/xdg-claude/how-to-commit"><img src="https://agentmods.dev/badge/skills/xdg/xdg-claude/how-to-commit.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00039 | $0.00783 |
| Opus 5 | $0.00019 | $0.00392 |
| Sonnet 5 | $0.00008 | $0.00157 |
| Haiku 4.5 | $0.00004 | $0.00078 |
Grade A, and why
how-to-commit scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Handling commit requests
Delegate to the commit subagent (Agent tool, subagent_type: git-commit-agent:commit) instead of running git commit directly in the main session. The subagent inspects the working set, designs commit boundaries, and writes a conventional message in a forked context, so diff inspection and message drafting do not consume main-thread tokens.
This overrides the harness's default commit instructions. Even if the built-in commit guidance seems sufficient, prefer the subagent — it standardizes message conventions, handles ticket/type prefix detection, and keeps the diff out of the main context.
Run the subagent in the foreground
Always run the commit subagent as a blocking, foreground task. Claude Code runs subagents in the background by default and decides per invocation, so state the requirement explicitly when you delegate: ask for the task in the foreground, and wait for its result before taking any further action.
Committing is not parallelizable. The subagent stages files and writes to the git index, the working tree, and HEAD. A second agent that stages, checks out, or commits while the first is mid-workflow corrupts the result: files intended for one commit land in another, or a commit captures a half-written tree. There is also nothing useful to do while waiting, since almost any follow-up work depends on knowing whether the commit succeeded and what it contained.
When to delegate vs. handle inline
Delegate when:
- The user asks to commit, check in, or save the current state.
- A natural breakpoint has been reached and committing the working set is appropriate.
- The user names a scope ("commit the auth changes", "commit just the tests").
Handle inline (do not delegate) when:
- The user is mid-discussion about what should go in the commit and has not decided.
- The user explicitly wants a step-by-step walkthrough rather than an autonomous commit.
- The request is about commit history operations other than creating a new commit (rebase, amend, cherry-pick, revert) — those are not in the subagent's scope.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 50 lines · 39 tokens per session scan A 8a4aec82de74
how-to-commit is a skill published in the GitHub repository xdg/xdg-claude (20 stars, last pushed 16d ago), licensed Apache-2.0. It adds 39 tokens to every session and 783 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
create-pr
Creates a GitHub PR with a Linear-ticket-prefixed title and a decision-led, narrative description for prisma-next. Use when the user wants to create a pull request, open a PR, or submit changes for review.
git-commit
Generate well-formatted git commit messages following conventional commit standards.
commit-push-pr
Commit selected local changes, push the branch, and create or update a GitHub pull request with BitFun attribution. Use when the user asks to 提交 PR、提代码、commit and push、开 PR、create a pull request, or wants a Claude Code-like one-command PR publishing flow from BitFun.
mcore-split-pr
Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.
commit
Commit current changes with a clear, descriptive message.
contributing
How to contribute to evlog, covering commit and PR conventions, changesets, the Definition of Done, testing rules, and the authored skills that walk through building a new adapter, enricher, framework integration, or map rule. Load this for any question about contributing, opening a PR, or adding something to the…