guild-audit

A static security checker for Guild plugin code. It reads hook scripts, tools, and optional MCP servers, then records hashes, declared access, and signs of network or filesystem use without running them.

In plain words
What is it for?
Use it to review Guild plugins before trusting them, check privacy and network access, and document the scripts and servers included in an installation.
Why use it?
It helps you inspect what installed plugin code can access and whether it may send data elsewhere. The dated report also lets you notice changes later by comparing file hashes.

Skill for Claude CodeCodex

Part of the guild plugin — 19 skills, 22 commands, 13 agents, 10 hooks, 2 MCP servers shipped together

Install

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.

agentmods
npx agentmods add skills/lookatitude/guild/audit
Any agent
npx skills add lookatitude/guild --skill audit
Clone the repo
git clone --depth 1 https://github.com/lookatitude/guild

Made for: Claude Code, Codex.

Or install guild, the plugin that ships this one along with the rest of its 19 skills, 22 commands, 13 agents, 10 hooks, 2 MCP servers.

Per session 92 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,846 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 3 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00092 $0.01846
Opus 5 $0.00046 $0.00923
Sonnet 5 $0.00018 $0.00369
Haiku 4.5 $0.00009 $0.00185

Measured 2d ago against content hash 0b70ec33e99c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

guild-audit scanned grade A with 3 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.

Unrestricted tool accesslowExcessive agency

A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.

The audit is **static** — it reads source files, computes hashes, greps for egress and filesystem patterns, and writes a dated report. It does not execute any script under audit.

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

- `network_evidence` — list of matched lines for `fetch(`, `http.request`, `https.request`, `WebFetch`, `WebSearch`, `axios`, `undici`, `node-fetch`, `got`, `curl` (in `.sh`). Each match records the file:line and the sur

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

- `filesystem_write_evidence` — list of matched lines for `writeFile`, `appendFile`, `createWriteStream`, `mkdir`, `rm`, `spawn(`, `exec(`, `execSync`, `>` / `>>` / `tee` in `.sh`. Each match records the file:line + the
skills/meta/audit/SKILL.md · 78 lines

How it starts

The opening of the file, as written. The whole thing — 78 lines — stays where its author put it; the contents beside it link to each section on GitHub.

guild:audit

Audits Guild plugin scripts for privacy and egress compliance: meta-skills are restricted to the filesystem; only the researcher has web access by default; /guild:audit surfaces script hashes. The plugin is only as safe as its installed-from source — install from trusted sources per Anthropic's guidance.

The audit is static — it reads source files, computes hashes, greps for egress and filesystem patterns, and writes a dated report. It does not execute any script under audit.

Scope

Audits every executable artifact that ships in the plugin:

  • hooks/*.ts and hooks/*.sh — SessionStart, UserPromptSubmit, PostToolUse, SubagentStop, Stop hooks per §13.2.
  • hooks/agent-team/*.tsTaskCreated, TaskCompleted, TeammateIdle hooks per §13.2.
  • scripts/*.ts — tooling scripts (trace-summarize, evolve-loop, flip-report, shadow-mode, description-optimizer, rollback-walker, and any others added in P7+).
  • mcp-servers/* — optional MCP servers per §13.3 (guild-memory, guild-telemetry), audited if present.

Does not audit:

  • skills/**/SKILL.md — markdown, no executable code. Skill bodies are instructions to the model, not programs; they are audited by the guild:evolve-skill gate and the guild:wiki-lint skill for content, not by this skill for egress.
  • agents/*.md — same rationale.
  • commands/*.md — same.
  • user application code — that's agents/security (the shipping specialist), not this skill.

Per-script report shape

One row per audited script, with the following fields:

  • path — absolute repo-relative path.
  • sha256 — hex digest of the file bytes.
  • loc — line count.
  • allowed_tools — from frontmatter when present (e.g. the hook has allowed-tools: declared), else n/a.
  • network_evidence — list of matched lines for fetch(, http.request, https.request, WebFetch, WebSearch, axios, undici, node-fetch, got, curl (in .sh). Each match records the file:line and the surrounding 1-line context. Empty list ⇒ no evidence of egress.
  • filesystem_write_evidence — list of matched lines for writeFile, appendFile, createWriteStream, mkdir, rm, spawn(, exec(, execSync, > / >> / tee in .sh. Each match records the file:line + the path argument literal when present.
  • writes_outside_guild — sub-list of the above where the path literal is NOT under .guild/runs/, .guild/evolve/, .guild/skill-versions/, .guild/reflections/, .guild/wiki/, .guild/team/, .guild/audit/, or a temp dir (os.tmpdir(), /tmp). Any non-empty value here is a red flag surfaced in the summary.
  • notes — free-form remarks (e.g. "uses child_process to shell out to git", "reads env var GUILD_TOKEN").

Read the full file on GitHub · 78 lines

Files

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.

Changes

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.

  1. 2d ago First seen · 78 lines · 92 tokens per session scan A 0b70ec33e99c

Subscribe to this mod's changes

guild-audit is a skill published in the GitHub repository lookatitude/guild (7 stars, last pushed 2d ago), licensed MIT. It adds 92 tokens to every session and 1,846 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 3 findings (unrestricted tool access, makes network calls, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

brainstorming

You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.

obra/superpowers · 37 tokens

auto-perf-optimize

Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.

microsoft/vscode · 62 tokens

chat-perf

Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.

microsoft/vscode · 51 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens