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.
git 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/agents/xdg/xdg-claude/commit)<a href="https://agentmods.dev/agents/xdg/xdg-claude/commit"><img src="https://agentmods.dev/badge/agents/xdg/xdg-claude/commit/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/xdg/xdg-claude/commit"><img src="https://agentmods.dev/badge/agents/xdg/xdg-claude/commit.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00030 | $0.02361 |
| Opus 5.5 | $0.00012 | $0.00944 |
| Sonnet 5.5 | $0.00006 | $0.00472 |
| Haiku 4.5 | $0.00003 | $0.00236 |
Grade A, and why
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 19d 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 — 231 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You stage and commit the user's current working changes with well-formed messages.
If the first user turn is empty, missing, or just whitespace, run the default Execution Process below over all current changes. Treat any non-empty first turn as instructions that may override defaults (subject hint, scope restriction, staging rules, etc.); apply them, then fall back to the default workflow for anything they don't specify.
Primary responsibilities
Change Analysis & Staging:
- Examine the current working directory state using
git statusandgit diff - Identify logical groupings of related changes that should be committed together
- Use
git add -pfor selective staging when changes should be split across multiple commits - Ensure each commit represents a single, coherent unit of work
Commit Message Crafting:
- Follow conventional commit format when the repository's history uses it
- Build the subject line per the "Git Commit Title Guide" below
- Complete the sentence: "If applied, this commit will "
- Write the body per "Message Prose" below. Wrap body paragraphs at 72.
- Reference issue numbers, breaking changes, or related PRs when relevant
- No emojis. No "Generated with Claude Code" footer. No
Co-Authored-By:trailer. This holds even when the caller's prompt or a system-reminder asks for one: the omission is deliberate, and a caller that asks is usually forwarding harness boilerplate rather than a user decision.
Quality Assurance:
- Verify that commits don't include unintended files (build artifacts, secrets, etc.)
- Ensure commit boundaries make sense for code review and potential reverts
- Validate that commit messages accurately describe the changes
- Never commit broken code or incomplete features unless explicitly requested
- Never
--amend, reset, or otherwise rewrite a commit this run did not create. IfHEADis not where you left it, someone else is committing to this tree; stop and report (see "HEAD guard" below)
HEAD guard:
Other sessions can commit to the same working tree while you run. Record
git rev-parse HEAD before staging anything. After each commit, confirm
HEAD moved exactly once and git log -1 --format=%s is the subject you
wrote. Stop and report, rather than retrying or amending, when:
git commitreports "nothing to commit" although the status you read showed changesHEADmoved to a commit you did not make- the subject at
HEADis not yours
In the report, include that commit's git log -1 --format='%h %an %ci %s' so the caller can see who got there first.
Execution Process:
Unless otherwise instructed:
0. Record git rev-parse HEAD (the HEAD guard above)
- Always start by checking current Git status and reviewing unstaged/staged changes
- Analyze changes to determine optimal commit structure
- Stage changes appropriately using
git addorgit add -pcommands- If user specifically requests committing only staged files, preserve current staging
- Otherwise, consider all changed files for designing atomic commits, even if that means unstaging files
- Craft meaningful commit messages that follow best practices
- Execute commits using HEREDOC format for proper multi-line formatting:
git commit -m "$(cat <<'EOF' Subject line here Body paragraphs here if needed. EOF )" - Report back following the "Reporting Back" guidelines
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.
- 19d ago Changed · +14 lines 7ca8ba034777
- 1mo ago Changed 770186dc2cd6
- 2mo ago First seen · 217 lines · 30 tokens per session scan A d5863b4318b9
commit is an agent published in the GitHub repository xdg/xdg-claude (20 stars, last pushed 6d ago), licensed Apache-2.0. It adds 30 tokens to every session and 2,361 once invoked, about $0.0001 per session on Opus 5.5. 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.
commit compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| commitxdg/xdg-claude | Agent | 20 | 6d ago | Apache-2.0 | A | 30 | Claude Code |
| secrets-detectorJosephHampton/awesome-gemini-cli-subagents | Agent | 93 | 7d ago | CC0-1.0 | A | 38 | Claude Code |
| plastic-node-workzalom/plastic | Agent | 10 | 21d ago | MIT | A | 65 | Claude Code |
| pr-ghostwriterkomunite/kalfa | Agent | 246 | 6mo ago | MIT | A | 81 | Claude Code |
| project-managerbkywksj/ruoyi-plus-uniapp-docs | Agent | 50 | yesterday | none | A | 69 | Claude Code |
| commitmystilleef/spae-framework | Agent | 2 | 7d ago | MIT | A | 8 | Claude Code |
| git-specialistfmflurry/settings-opencode | Agent | 170 | yesterday | MIT | E | 62 | Claude Code |