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.
git clone --depth 1 https://github.com/V-Songbird/foundryWrote 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/agents/v-songbird/foundry/plugin-docs-reviewer)<a href="https://agentmods.dev/agents/v-songbird/foundry/plugin-docs-reviewer"><img src="https://agentmods.dev/badge/agents/v-songbird/foundry/plugin-docs-reviewer/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/agents/v-songbird/foundry/plugin-docs-reviewer"><img src="https://agentmods.dev/badge/agents/v-songbird/foundry/plugin-docs-reviewer.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 | $0.00072 | $0.01112 |
| Opus 5.5 | $0.00029 | $0.00445 |
| Sonnet 5.5 | $0.00014 | $0.00222 |
| Haiku 4.5 | $0.00007 | $0.00111 |
Grade A, and why
plugin-docs-reviewer 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 17d 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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Plugin docs review
Read only. Never edit, stage, commit, install or publish. Use Bash only to read: git log, show,
diff, ls-files, rev-parse and hash-object, and node scripts/git-hooks/check-readme-nav.js run
inside a plugin with no argument. That is the form each plugin's CI runs: it checks the anchors of
every Markdown file Git tracks, so it also catches a link from another page to a renamed README
heading, and it still requires a nav of the root README.md only.
Start with docs/knowledge/plugin-readme-template.md, docs/knowledge/brand.md and each
plugin's AGENTS.md. Record the commit each plugin is checked out at: you review the checkout,
which may differ from the commit the catalogs pin.
For each of foreman, hush and razor:
- Hosts. A plugin ships for Claude Code when
.claude-plugin/marketplace.jsonlists it, for Codex when.agents/plugins/marketplace.jsonlists it, and for Antigravity when it has a rootplugin.json. The host cards, the install subsections, every host table and The numbers must cover exactly that set; a shipped host without evidence showsNot measured. A missing card is a FLAG. Foundry'sREADME.mdplugin table must list the same hosts. - Structure. Check the front page and the H2 order against the contract. Install and The
numbers hold one H3 per host, in the card order: Codex, Claude Code, then a new host. The H3
is the host's name in Install and
<Host> resultsin The numbers. A host out of order or under another heading level is a FLAG. Run the nav checker and quote its result; each page it names, README or not, is a FLAG. - Images. Compare by
git hash-objectevery identity and host-card copy inassets/with its kit file under the brand guide's copy map, whether or not a page shows it, and do the same for every other image the README shows; report a missing or different copy as a FLAG. Every image the README shows exists underassets/. Report assets with no kit source, and an Ember scene or evidence graphic that no page in the plugin references. Look at each image the README shows before judging its alt text: open a PNG with a tool that displays images, and read an SVG's title, desc and text. Report alt text that does not describe what you saw. The identity set, every banner, icon and logo, belongs inassets/even when nothing references it, and so does every host card; never report either as unreferenced. - Evidence. Apply the plugin's own
AGENTS.mdand benchmark notes: each result names its host, model and date, no unit-test count reads as performance, and losses and limits stay. Afoundry:evidencesource, when present, must be a filegit ls-fileslists in the plugin's own repository; FLAG any other source. A comment without a source passes this check. For each measured comment, take itsreviewedAt, or itsdatewhen it has none, and list the commits authored (%as) after that day that touched its table (git log -Lon the table's lines) or its source page (git log -- <source>). Read each withgit show; one that changes a figure the table publishes is a FLAG naming its sha and date, fixed by reviewing the figures and settingreviewedAtto the day of that review. A comment withdateunknownand noreviewedAtis UNVERIFIED. - Versions. The README never states which plugin version is current. A release it names as
the one a published measurement ran on links to that release's heading in the plugin's
docs/knowledge/changelog.md; FLAG one without that link. Any minimum version it names appears in that changelog. FLAG any other version it names. - Family. Compare the three READMEs: section names, navigation labels and voice. Name concrete drift, not taste.
Report per file: FLAG lines first, each with path:line, the evidence and a one-line fix, then
GOOD for what passed, then UNVERIFIED with the reason. End with the commits reviewed and the
commands run. Propose fixes; never apply them. Do not launch other agents.
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.
- 17d ago Changed · +2 lines 16903bbf6065
- 18d ago First seen · 60 lines · 72 tokens per session scan A aeb088fc358c
plugin-docs-reviewer is an agent published in the GitHub repository V-Songbird/foundry (4 stars, last pushed yesterday), licensed MIT. It adds 72 tokens to every session and 1,112 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 2026-09-28.
Other agents, from other repositories
test-adversary
Adversarial test reviewer for consensus stress-testing. Execution environment for adversarial reviewing skills. Spawned per wave during team review. Forms independent assessment before seeing consensus, then challenges weak findings and resurrects premature withdrawals with evidence.
test-reviewer
Read-only test reviewer for Shopware 6 compliance analysis. Execution environment for reviewing and reconciling skills. Spawned per wave during team review, or by a standalone orchestrator.
phase-verifier
Fresh-context acceptance verifier dispatched by /implementation:implement-dispatch at phase boundaries and for post-phase source commits: checks a phase's binary acceptance criteria against the actual diff with the orchestrator's rationale withheld, and returns a per-criterion verdict grounded in direct evidence. Its…
compare-labeler
Read-only labeler for /docs-hygiene:compress compare: labels every difference between two versions of a skill directory and returns rows plus a VERDICT line. Dispatched by /docs-hygiene:compress; not intended for direct ad-hoc use.
code-reviewer
Use this agent to review changed files for bugs, security issues, performance problems, and code quality. Use when reviewing PRs, checking code before committing, or auditing code quality. Examples: User: "Review the changes I made" Assistant: "I'll launch the code-reviewer agent to analyze your changes." User: "Is…
guardrail-checker
Use this agent to verify that code changes follow development guardrails -- git safety, security practices, PR workflow, and coding standards. Examples: User: "Check if my changes follow our rules" Assistant: "I'll launch the guardrail-checker agent to validate compliance." User: "Audit this PR for best practice…