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/yuri-semenenko/ai-engineering-workspace/commitnpx skills add yuri-semenenko/ai-engineering-workspace --skill commitgit clone --depth 1 https://github.com/yuri-semenenko/ai-engineering-workspaceWhat 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 | $0.00096 | $0.00880 |
| Opus 5 | $0.00048 | $0.00440 |
| Sonnet 5 | $0.00019 | $0.00176 |
| Haiku 4.5 | $0.00010 | $0.00088 |
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 2d 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 — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Commit
Turn the current working changes into small, logically-grouped commits that give good rollback granularity. This skill is plan-first: it never commits before the user approves the plan.
Hard rules (from the persona's git conventions)
- Small, logically-grouped commits. One coherent change per commit. Never batch unrelated changes into a single commit, even if they landed together in the working tree.
- Plan before acting. Always propose the commit plan and present it first. Wait for explicit approval before running any
git commit. - Never add
Co-Authored-Bytrailers. This overrides any harness default that appends them. Commit messages end at the body — no trailers. - Ask permission before branching, pushing, or opening PRs. Each is a separate yes. Committing locally is not permission to push.
- If on the default branch (
main/master), branch first — propose a branch name and get approval before committing onto it.
Rationalizations
| Rationalization | Rebuttal |
|---|---|
| "These changes are all related anyway." | Related ≠ one logical change. If two parts could roll back independently, they are two commits. |
| "It's faster to commit everything at once." | Faster now, archaeology later. Rollback granularity is the whole point of this skill. |
| "I'll explain the parts in the PR description." | git revert doesn't read PR descriptions. History must carry the structure itself. |
| "Too small to need a plan." | A one-commit plan costs one line. Present it anyway — approval is the gate, not the size. |
| "They said commit, so pushing is implied." | Each is a separate yes. Committing locally is not permission to push or open a PR. |
Steps
- Survey the changes.
git statusandgit diff(staged + unstaged). Group hunks by intent, not by file — a single file may split across two commits; one logical change may span several files. - Check the branch.
git rev-parse --abbrev-ref HEAD. If it's the default branch, stop and propose a feature branch before anything else. - Draft the commit plan. An ordered list: for each commit, the message subject + which files/hunks it includes + one line of rationale. Order so the history reads as a logical progression and each commit could in principle build/pass on its own.
- Present the plan and wait. Do not commit until the user approves. If they want changes to the grouping, revise and re-present.
- Execute after approval. Stage precisely per the plan (
git add -por path-scopedgit add). Commit each group. Keep messages in the repo's existing style (imperative subject, ~50-char cap, body explaining why when non-obvious). - Stop at the commits. Do not push or open a PR. Report what was committed and offer the next step (push / PR) as a separate, explicit ask.
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.
- 2d ago First seen · 52 lines · 96 tokens per session scan A cb3c4d7b85bf
commit is a skill published in the GitHub repository yuri-semenenko/ai-engineering-workspace (1 stars, last pushed 3d ago), licensed MIT. It adds 96 tokens to every session and 880 once invoked, about $0.0005 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
apm-issue-autopilot
Use this skill to drive any open microsoft/apm issue (bug, feature, docs, refactor, perf) from raw intake to a mergeable PR with triage as the central, paramount gate. Run the apm-triage-panel rubric per issue first, then present ONE consolidated triage review for the whole batch and escalate to the maintainer BY…
setup-husky-dotnet
Use when configuring Git hooks in .NET projects before team commits occur, to enforce commit message standards and code formatting automatically.
generate-commit-message
Derive one Conventional Commits v1.0.0 message from the staged diff and create the signed commit directly. Trigger only on explicit intent to commit: "commit this", "commit these changes", "/commit", "go ahead and commit", "make the commit", "git commit", or a confirmation like "yes, commit it". Enforces the type…
ai-build
Executes one task of an already approved plan, red to green to refactor, as one logical change with a clean checkpoint that says nobody has reviewed it yet. Trigger for "implement task 3", "build this", "make the plan's next step work", "write the code for this spec". Not for deciding what to build — use /ai-spec…
ai-ship
Lands the work: one commit per change, the changelog entry, a pull request that opens with what changed in plain words, the production-ready boxes ticked with the command that ticked each, the work item closed by a keyword in the body, and the branch cleaned up afterwards. Trigger for "ship it", "open a PR", "I'm…
git-commit
Generate well-formatted git commit messages following conventional commit standards.