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/outlinedriven/odin-claude-plugin/setup-pre-commitnpx skills add OutlineDriven/odin-claude-plugin --skill setup-pre-commitgit clone --depth 1 https://github.com/OutlineDriven/odin-claude-pluginWrote 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/outlinedriven/odin-claude-plugin/setup-pre-commit)<a href="https://agentmods.dev/skills/outlinedriven/odin-claude-plugin/setup-pre-commit"><img src="https://agentmods.dev/badge/skills/outlinedriven/odin-claude-plugin/setup-pre-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 | $0.00071 | $0.01449 |
| Opus 5 | $0.00036 | $0.00724 |
| Sonnet 5 | $0.00014 | $0.00290 |
| Haiku 4.5 | $0.00007 | $0.00145 |
Grade A, and why
setup-pre-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 — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Setup pre-commit
Contract
| Field | Bound contract |
|---|---|
| Trigger | Install, repair, or standardize local pre-commit checks in an existing repository, including package-manager-native commit-time checks. |
| Authority | Reversible local writes to hook configuration, dependency manifests, and lockfiles. No push, remote setting, credential, or release mutation. |
| Side effect | Extends the existing hook manager or installs one current manager, writes its config, and installs the local Git hook. |
| Done | Exactly one hook manager owns pre-commit; its all-files command runs the repository formatter, linter, type checker, and targeted tests successfully; a deliberate failing probe blocks and a passing probe exits zero. |
Inputs
- Repository root with
.git: required. - Repository-native format, lint, type-check, and targeted-test commands: required; infer only from committed scripts and configuration.
- Existing hook manager: optional; detect Husky, Lefthook, prek/pre-commit, cargo-husky, and native
.git/hooks/pre-commitbefore adding anything.
Procedure
- Read the repository manifests, lockfiles, declared gate scripts, hook config, and
.git/hooks/pre-commit. Detectpnpm-lock.yaml,bun.lock, legacybun.lockb,uv.lock,go.mod,Cargo.toml, anddune-project. Do not infer a gate command that the repository does not declare. Done when: the repository's manifests, lockfiles, gate scripts, and existing hooks are inventoried. - If one hook manager already exists, extend it. If multiple managers can fire for the same commit, stop and report the conflict; do not add a third path. Done when: the hook-manager state is classified as extend, conflict, or install-new.
- If none exists, select one manager: Lefthook for a pnpm, Bun, or Go repository; prek for Python, Rust, or OCaml. In a mixed repository, select the manager already represented by its lockfile or task runner. Ask only when two choices remain equally supported by repository evidence. Done when: the manager is selected or the user is asked.
- Install through the current project toolchain:
pnpm add -D lefthook @biomejs/biome && pnpm exec lefthook installfor JavaScript or TypeScript;go install github.com/evilmartians/lefthook@latest && lefthook installfor Go;uv tool install prek && prek installfor Python, Rust, or OCaml. Do not install ESLint, Prettier, Black, isort, mypy, or a second package manager. Done when: the manager is installed. - Write only commands the repository can execute:
- JavaScript or TypeScript Lefthook:
pnpm exec biome check --write --no-errors-on-unmatched ., the declared type-check script, and the declared targeted-test script. - Python prek local hooks:
uv run ruff check --fix .,uv run ruff format --check .,uv run pyright, and the declared targeted pytest command. - Go Lefthook: fail when
gofmt -l .returns a path, then rungo vet ./...andgo test ./.... - Rust prek local hooks:
cargo fmt --check,cargo clippy --all-targets -- -D warnings, and the repository test command, preferringcargo nextest runwhen configured. - OCaml prek local hooks:
dune fmt,dune build @runtest, and any stricter repository alias already declared. Setpass_filenames: falsefor whole-repository commands. Keep independent read-only checks parallel only when their tools do not edit the same files. Done when: the config is written with repository-native commands only.
- JavaScript or TypeScript Lefthook:
- Run the manager's all-files entry point:
pnpm exec lefthook run pre-commitorprek run --all-files. If a formatter changes files, inspect the diff and repeat until the non-mutating gate passes. Done when: the all-files command passes with no uncommitted formatter changes. - Prove enforcement without creating history. Add one temporary, reversible formatting violation inside an owned scratch file that the hook includes. Run the hook and require non-zero or an automatic repair followed by a dirty diff. Restore the scratch file, rerun the hook, and require zero. When the hook does not include a formatter, use the trivial-commit alternative: stage a trivial change (e.g., add a blank line to a tracked file), run the hook, and require all configured checks to pass; then revert the staged change. Never weaken a command to make the probe pass. Done when: the failing probe blocks and the passing probe exits zero, or the trivial-commit probe passes all checks and is reverted.
- Confirm exactly one executable
.git/hooks/pre-commitentry remains and that it delegates to the selected manager. Report changed files and both probe outcomes. Do not commit unless the user separately asks. Done when: one hook entry is confirmed and the report is emitted.
What ships with it
2 files 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 · 52 lines · 71 tokens per session scan A 0a00a4ba3c43
setup-pre-commit is a skill published in the GitHub repository OutlineDriven/odin-claude-plugin (35 stars, last pushed yesterday), licensed Apache-2.0. It adds 71 tokens to every session and 1,449 once invoked, about $0.0004 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 2026-09-04.
Other skills, from other repositories
java-commit
Generates a Conventional Commits message for staged Java changes. Use when user asks to "write a commit message", "help me commit", "what should my commit say", "summarize my changes", "draft a commit", or "create commit message".
claude-code-bash-patterns
Claude Code Bash tool patterns with hooks, automation, git workflows. Use for PreToolUse hooks, command chaining, CLI orchestration, custom commands, or encountering bash permissions, command failures, security guards, hook configurations.
slop-scan
Use when scanning tracked repository files for AI slop, verbosity, brittle references, or low-value contributions.
worktrees
Use when fanning several tickets into parallel branches and git worktrees, with related tickets grouped together.
weave-fix-review
Fix weave review findings — validate, add test coverage, fix, and commit each as atomic changes.
action
Use when screened review findings should be fixed on the current branch, usually as one verified commit per finding.