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 jlifyio/wyx --skill auditgit clone --depth 1 https://github.com/jlifyio/wyxWrote 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/jlifyio/wyx/audit)<a href="https://agentmods.dev/skills/jlifyio/wyx/audit"><img src="https://agentmods.dev/badge/skills/jlifyio/wyx/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/jlifyio/wyx/audit"><img src="https://agentmods.dev/badge/skills/jlifyio/wyx/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 | $0.00081 | $0.01706 |
| Opus 5.5 | $0.00032 | $0.00682 |
| Sonnet 5.5 | $0.00016 | $0.00341 |
| Haiku 4.5 | $0.00008 | $0.00171 |
Grade A, and why
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 1mo 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 — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Audit & Command Planner
Scan project for wyx spec coverage and generate a prioritized action plan. This skill does NOT generate specs — it identifies what needs specs and outputs commands to run. Unlike SessionStart (which reports counts), this adds: pattern-based pipeline/sync candidate detection and dependency-ordered command sequences.
Tool constraint: Prefer Glob, Grep, and Read for all discovery and counting; do NOT
use subagents. Degraded mode — Claude Code leaves Glob and Grep out by default on
macOS, Linux and WSL, so expect this path there; when they are unavailable, fall back to read-only shell via Bash for discovery only
(find / ls / grep -r to list and read files). Never use Bash to write, edit,
move, or generate files — audit stays strictly read-only on every
discovery path: it reports, it never produces artifacts.
How to interpret $ARGUMENTS
- Path (e.g.
src/lib/server/): Scope the audit to this subtree only. - No arguments: Audit the entire project.
Step 1: Discover Existing Specs
Glob for all wyx spec files (scoped to path if given):
**/CONCEPT.md**/PIPELINE.md**/SYNCS.md
Read each spec's first heading to extract the concept/pipeline/sync name.
For each spec, also check for boundary-contributing sections (the sections the PreToolUse
hook extracts): CONCEPT.md should have ## interactions or ## dependencies; PIPELINE.md
should have ## data boundary. Note any spec missing all its expected boundary sections.
Step 2: Identify Uncovered Modules
Find directories with more than 2 source files lacking any colocated spec (CONCEPT.md, PIPELINE.md, or SYNCS.md). This matches the SessionStart hook's exclusion set — any of the three spec types protects a directory from being flagged.
Prefer Glob; if no Glob tool exists in the harness, use read-only shell (find / ls) for discovery only — never to write or modify files.
How to count source files per directory using Glob:
- Run Glob for
**/*.{ts,js,tsx,jsx,svelte,vue,py,rs,go,java}(scoped to path if given) - From the results, group files by their parent directory and count per directory
- Directories with more than 2 source files and no colocated spec are candidates
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.
- 1mo ago Changed 68bd592adcd6
- 2mo ago First seen · 162 lines · 81 tokens per session scan A 28a7c842bb27
audit is a skill published in the GitHub repository jlifyio/wyx (3 stars, last pushed 29d ago), licensed MIT. It adds 81 tokens to every session and 1,706 once invoked, about $0.0003 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 2026-08-31.
audit compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| auditjlifyio/wyx | Skill | 3 | 29d ago | MIT | A | 81 | Claude Code |
| architecture-auditkarsten-s-nielsen/mad-scientist-skills | Skill | 3 | 1mo ago | MIT | A | 174 | Claude Code |
| final-reviewkarsten-s-nielsen/mad-scientist-skills | Skill | 3 | 1mo ago | MIT | A | 80 | Claude Code |
| unbiased-reviewkarsten-s-nielsen/mad-scientist-skills | Skill | 3 | 1mo ago | MIT | A | 84 | Claude Code |
| refactorzircote-plugins/refactor | Skill | 2 | 5mo ago | MIT | A | 112 | Claude Code |
| releasehyhmrright/brooks-lint | Skill | 1.6k | yesterday | MIT | A | 135 | Claude Code, Codex |
| recipe-front-reviewshinpr/claude-code-workflows | Skill | 711 | 4d ago | MIT | A | 29 | Claude Code |