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/doc-reviewnpx skills add OutlineDriven/odin-claude-plugin --skill doc-reviewgit clone --depth 1 https://github.com/OutlineDriven/odin-claude-pluginWhat 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 | $0.00032 | $0.04047 |
| Opus 5 | $0.00016 | $0.02024 |
| Sonnet 5 | $0.00006 | $0.00809 |
| Haiku 4.5 | $0.00003 | $0.00405 |
Grade A, and why
doc-review 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 — 248 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Doc-review: multi-persona content-shape review of plans and specs
doc-review evaluates a requirements doc, plan, spec, or PRD through specialist reviewer lenses. It classifies the document by content shape, selects the lenses the document actually warrants, dispatches them in parallel as read-only subagents, synthesizes their confidence-anchored findings through a multi-stage pipeline, and routes each survivor into one of four handling tiers. The structural invariant: it never edits the document under review. The only surface it may write is a single review-record file, and only when persistence is requested.
Auto-invoke
<auto_invoke> <trigger_phrases>
- "review this plan"
- "review this spec"
- "review this PRD"
- "review these requirements"
- "critique this design doc"
</trigger_phrases>
Fire automatically on a trigger phrase against a prose document, or on
/doc-review. Auto-firing is permission to evaluate, not permission to fabricate. Every finding still has to clear the evidence-quote and confidence-anchor floor below. A clean "nothing above the floor" is a valid, correct result. <manual_override>/doc-review [path]reviews the named document, or lists candidates to choose from when no path is given.--recordpersists a review-record file and stages only it.mode:headlessmakes the run non-interactive (return structured findings, no questions, no record unless--record).</manual_override> </auto_invoke>
When to Apply / NOT
Apply when the user wants a prose planning document evaluated: a requirements doc, a plan, a spec, a PRD, a design doc, a brainstorm. The deliverable is findings (and, on request, one review-record file).
NOT source code: if the target is a diff or code file, stop and route to review. This skill does not open a diff.
Support files: read on demand
Don't bulk-load these at start. Read each file only when the workflow phase that needs it fires. Each file is self-contained.
Persona files (read when that persona is selected; paste full content into that subagent's prompt):
What ships with it
14 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.
- agents/openai.yaml 109 B
- references/bulk-preview.md 7.1 KB
- references/findings-schema.json 4.2 KB
- references/open-questions-defer.md 3.8 KB
- references/personas/adversarial.md 4.5 KB
- references/personas/coherence.md 3.1 KB
- references/personas/feasibility.md 3.4 KB
- references/personas/product.md 3.9 KB
- references/personas/scope-guardian.md 3.4 KB
- references/personas/security.md 3.2 KB
- references/review-output-template.md 8.8 KB
- references/subagent-template.md 26 KB
- references/synthesis-and-presentation.md 26 KB
- references/walkthrough.md 9.7 KB
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 · 248 lines · 32 tokens per session scan A 76d987dae3ac
doc-review is a skill published in the GitHub repository OutlineDriven/odin-claude-plugin (35 stars, last pushed 5d ago), licensed Apache-2.0. It adds 32 tokens to every session and 4,047 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
cass
Mine past agent sessions for working prompts, decisions, and patterns (session archaeology). Triggers: "cass", "mine past agent sessions for", "cass skill".
codebase-recon
Reconstruct a repository as cited entry-to-test flows and bounded claims. Triggers: "codebase recon", "trace this codebase", "repository audit", "refresh the prior recon".
craft-goal
Compile or lint a persistent Mayor-style goal prompt that ratchets a bead graph through bounded RPI experiments toward one larger outcome. Triggers: "craft a goal prompt", "mayor goal", "goal-runner prompt", "lint this goal", "is this goal safe". (Shaping one experiment's intent routes to plan.).
ms
Skill "ms" from boshu2/agentops, covering ms — metaskill search/load engine, constraints, quick start, consume — mcp-primary (mcpms) and write / admin — cli-only (verified landing in the live db).
reverse-engineer
Reverse-engineer an authorized repo, binary, or product into a verifiable feature inventory and adoption map. Triggers: "reverse-engineer X", "tear down Y", "what should we steal from Z", "evaluate competitor/upstream", "should we fork/adopt/build-native".
skill-eval
Author and tier behavioral probes for a skill, including seeded-defect probes that escape ceiling saturation. Triggers: "measure this skill", "the probe came back INERT", "the control arm aces it", "harden this scenario", "is this skill actually doing anything".