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/thettwe/nyann/commitnpx skills add thettwe/nyann --skill commitgit clone --depth 1 https://github.com/thettwe/nyannWhat 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 | $0.00175 | $0.02182 |
| Opus 5 | $0.00088 | $0.01091 |
| Sonnet 5 | $0.00035 | $0.00436 |
| Haiku 4.5 | $0.00017 | $0.00218 |
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 yesterday.
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 — 215 lines — stays where its author put it; the contents beside it link to each section on GitHub.
commit
You are generating a commit message for the user's staged changes. Work
in phases. Never run git commit until the user has confirmed the message.
0. Drift check (quick, non-blocking)
Run bash bin/session-check.sh --flow=commit. If it produces output,
surface the line to the user verbatim. Do not block the flow.
1. Gather context
Run bin/commit.sh --target <cwd>. The script emits a JSON context object
(target, branch, on_main, convention, staged_files, summary,
diff, truncated).
Exit code handling:
0→ JSON context emitted. Check thenothing_stagedfield:nothing_staged: true→ ask the user what they want to stage (git add <paths>,git add -p, or "everything"), then re-run.- Otherwise → full context available. Continue.
2→ not a git repo. Tell the user, stop.
If on_main: true, warn the user that the block-main hook will reject
the commit. Suggest they create a feature branch via the new-branch
skill before continuing.
1.5. Pre-action guards + commit hygiene
Run both of these before generating the message — they inform scope choice and prevent obvious mistakes from reaching the commit message step:
bash bin/pre-action-guard.sh --flow commit --target <cwd> [--profile <resolved-profile.json>]
bash bin/commit-hygiene.sh --target <cwd> [--profile <resolved-profile.json>]
Guard handling (exit code)
| Exit | Meaning | Action |
|---|---|---|
| 0 | All guards passed (or only advisory warnings) | Surface any advisory message lines, continue |
| 3 | One or more critical guards failed | Surface the failing guard messages. Refuse to proceed unless the user passes --skip-guards explicitly (re-ask via AskUserQuestion). |
| 4 | One or more confirm-severity guards failed |
Surface the failures. Call AskUserQuestion: "Proceed despite the warnings?" Only continue on explicit confirm. |
Don't silently pass critical failures. The merge-conflict-markers
guard catching <<<<<<< in the staged diff is exactly the case where
"warn and continue" would let a broken merge into a commit.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- yesterday First seen · 215 lines · 175 tokens per session scan A 389cc6651d1c
commit is a skill published in the GitHub repository thettwe/nyann (6 stars, last pushed 2d ago), licensed MIT. It adds 175 tokens to every session and 2,182 once invoked, about $0.0009 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 2026-08-31.
Other skills, from other repositories
github-skill-forge
一个"制造技能的技能"。这个工具自动化了将任意 GitHub 仓库转换为标准化 Trae 技能的全过程，是扩展 AI Agent 能力的核心工具。.
issue
Use when starting a chain from a GitHub issue — turning an issue URL or number into a triaged, planned, dispatched, and reviewed pull request. Classifies the thread (bug → root-cause discipline, feature → plan chain, question → drafted reply), synthesizes a spec from the issue's own acceptance criteria, then runs the…
github
GitHub operations via gh CLI: issues, PRs, CI runs, code review, API queries.
silicon-throne
硅基圣座总控核心，以神谕式统御把零散任务压成 末将-first 之文明工程.
github-image-upload
Upload local images and other files (PDF, zip, log, …) to GitHub and embed them in a pull request description, an issue, or a comment — producing canonical github.com/user-attachments URLs (private-repo uploads stay private). Use when asked to "attach a screenshot to the PR", "add an image to the PR description", "put…
a11y-remediate
Use to produce a leader-facing remediation proposal from one or more /a11y-audit outputs plus team and product context. Translates audit findings into sprint plans, staffing asks, customer-facing language, compliance rollups, and critical-path analysis. Refuses to fabricate numbers, owners, or commitments beyond the…