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 skills add MithunWijayasiri/dotclaude --skill skill-auditgit clone --depth 1 https://github.com/MithunWijayasiri/dotclaudeWrote 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/mithunwijayasiri/dotclaude/skill-audit)<a href="https://agentmods.dev/skills/mithunwijayasiri/dotclaude/skill-audit"><img src="https://agentmods.dev/badge/skills/mithunwijayasiri/dotclaude/skill-audit/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/skills/mithunwijayasiri/dotclaude/skill-audit"><img src="https://agentmods.dev/badge/skills/mithunwijayasiri/dotclaude/skill-audit.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 | $0.00045 | $0.01091 |
| Opus 5.5 | $0.00018 | $0.00436 |
| Sonnet 5.5 | $0.00009 | $0.00218 |
| Haiku 4.5 | $0.00005 | $0.00109 |
Grade A, and why
skill-audit 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 10d 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 — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill Audit
Report-first. Edit only approved findings.
Setup
- Load
compact-markdown; read itsskill-authoring.mdreference — that is the rubric. Don't restate it in the report or here. - Resolve targets:
- Path to a folder → use it. Path to
SKILL.md→ its parent folder. - Name →
<repo>/.claude/skills/<name>/, then~/.claude/skills/<name>/. Both exist → ask which. - No args → list skills in both locations, ask which.
- Path to a folder → use it. Path to
Read
Per target:
SKILL.mdand every file under the skill folder (references/, scripts).- What it delegates to or cites: subagent defs (
.claude/agents/,~/.claude/agents/), sibling skills it routes to, rules and CLAUDE.md files it references. - Every other skill's name + description: both skill locations plus plugin skills in the session's skill listing. Feeds the collision check.
Checks
Rubric from skill-authoring.md: description shape, triggering, ladder, failure modes (premature completion, duplication, sediment, sprawl, no-op). Rationale lines are content — don't flag as filler.
Audit-only additions:
- Stale refs — every cited path, file, agent, skill, command, script, and folder exists. Verify with Glob/Grep, never from memory. Missing → finding; unclear replacement → ask, don't guess.
- Frontmatter conflict —
disable-model-invocation: truealongside auto-trigger prose ("when the user asks…", "do NOT trigger for…"). - Instruction conflict — contradicts global/project CLAUDE.md or a rule (e.g. routes to a skill/agent CLAUDE.md doesn't sanction, auto-runs checks CLAUDE.md forbids).
- Cross-file duplication — restates its subagent's contract, CLAUDE.md, or a rule → pointer or cut.
- Collision — triggers overlap another skill or a built-in slash command, with no
Not for → <sibling>clause on either side. Manual-only → name collision only. - Approval gate — step that deletes, pushes, writes externally (Jira, PR, message), or overwrites, with no ask-first step before it.
- Scope fit — body does more than the description promises, or
allowed-toolsdoesn't match the tools the steps use. - Failure path — step that can fail (tool missing, command errors, ref not found) with no stated next action. Silent continue = finding.
- Description — see below.
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.
- 10d ago Changed · -11 tokens per session 7d52a21262c4
- 11d ago Changed · +2 lines eb34dd3d2d03
- 15d ago First seen · 73 lines · 56 tokens per session scan A 0b9c69973842
skill-audit is a skill published in the GitHub repository MithunWijayasiri/dotclaude (4 stars, last pushed 6d ago), licensed MIT. It adds 45 tokens to every session and 1,091 once invoked, about $0.0002 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 2026-10-06.
Other skills, from other repositories
auditing-skills
Use when checking this repository's skills for security or quality issues before sharing them, or remediating findings across skills.
ship-it-or-fix-it
Oracle-frozen Builder and independent-Judge convergence cycle. Load ONLY when the operator explicitly sets Governance Dial G2 for the task, or explicitly names this skill or an active work unit already running it. Never auto-activate on task class, such as security, auth, or payments. If a task seems to warrant G2 and…
test-verification
Requires behavioral, failure-path, and durable-seam evidence for tests and review. Use when writing tests, reviewing test coverage, assessing behavioral test quality, or accepting high-risk behavior on test evidence.
nitpicker
Hostile audit toolkit: one entry point dispatching specialist commands — adversarial review, security, tests, docs, types, architecture, performance, reliability, caching, concurrency, error handling, resource leaks, dependencies, licensing, CI, commits, migrations, observability, API contracts, a11y, i18n, privacy…
new-command
Scaffolds a new nitpicker audit command through the full RED → GREEN → REFACTOR → adversarial-review → validate cycle, wiring its command file, dispatch-table row, and docs together. Use when adding a new /nitpicker command to this repository, or when asked to create, scaffold, or add an audit command here.
pr-buddy
Two-phase PR review companion. Phase 1: Socratic conversational walkthrough that builds the reviewer's mental model through guided questions — fighting cognitive debt by ensuring comprehension before approval. Phase 2: automated quality checks via pr-review-toolkit (Claude Code) or inline review (any host). Use when…