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 HirogaKatageri/hirokata --skill research-peer-reviewgit clone --depth 1 https://github.com/HirogaKatageri/hirokataWrote 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/hirogakatageri/hirokata/research-peer-review)<a href="https://agentmods.dev/skills/hirogakatageri/hirokata/research-peer-review"><img src="https://agentmods.dev/badge/skills/hirogakatageri/hirokata/research-peer-review/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/hirogakatageri/hirokata/research-peer-review"><img src="https://agentmods.dev/badge/skills/hirogakatageri/hirokata/research-peer-review.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 | $0.00144 | $0.00987 |
| Opus 5 | $0.00072 | $0.00494 |
| Sonnet 5 | $0.00029 | $0.00197 |
| Haiku 4.5 | $0.00014 | $0.00099 |
Grade A, and why
research-peer-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 11d 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 — 93 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research Peer Review — Audit Before You Rely
Every confident research output hides weak claims, quiet biases, and missing angles. STORM's own documented weakness is that it doesn't self-critique — so source bias and fact misassociation (a real source bolted to a claim it doesn't actually make) slip through. This skill is the fix, usable on any research artifact, not just STORM output. Real peer review takes months; this runs an honest, adversarial version in minutes.
What It Checks
- Fact-check & hallucination audit — unsupported, exaggerated, or fabricated claims; citations that don't support what they're attached to; spot-checked against live sources.
- Bias detection — over-/under-represented views; default tilt toward mainstream/academic framing.
- Completeness — missing angles, stakeholders, counterexamples, recent developments.
- Contradiction handling — was dissent represented fairly, or quietly buried? Are reliability ratings honest?
- Actionability & clarity — are conclusions specific and justified, or vague and overstated?
Output: a dimension-by-dimension assessment, a ranked list of required fixes, and an overall reliability grade (A–F).
Workflow
Step 1 — Locate the Artifact
Identify what to review:
- STORM briefing:
.storm/{slug}/briefing.md(with its perspective files and contradiction map as supporting evidence). - A document/report: the user names a path or pastes the content.
- An argument in the conversation: the claims are already in context.
Establish a topic label and, for file-based runs, the workspace .storm/{slug}/ for the output.
If the target is ambiguous, ask one question to pin it down.
Step 2 — Spawn the Peer Reviewer
Launch the peer-reviewer agent with one Task call:
"Topic: {topic}. Review {artifact path OR pasted content} across all five dimensions. Supporting evidence (if present): {perspective files, contradiction map}. Spot-check the highest-stakes claims with WebSearch/WebFetch before flagging. Write
.storm/{slug}/peer-review.mdwith a reliability grade and ranked required fixes."
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.
- 11d ago First seen · 93 lines · 144 tokens per session scan A 98817e6d6244
research-peer-review is a skill published in the GitHub repository HirogaKatageri/hirokata (5 stars, last pushed 3d ago), licensed MIT. It adds 144 tokens to every session and 987 once invoked, about $0.0007 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 2026-08-31.
Other skills, from other repositories
unslop-review
Rewrites code review comments so they read like a human teammate wrote them. Cuts corporate-AI throat-clearing ("I noticed...", "I was wondering if perhaps...", "It might be worth considering..."). Each comment is direct: location, the issue, a concrete fix. Use when user says "humanize review", "de-slop PR comment"…
code-review
A code-review procedure that checks changed code by running available tests and type checks, then examining correctness, security, performance, style, and maintainability.
autoreview
Pre-commit/ship code review: Codex default; optional Claude or Pi.
omh-code-review
This is a Hermes-native code-review workflow skill.
revdiff-plan
Review the last Codex assistant message (plan, analysis, or proposal) with inline annotations in a TUI overlay. Extracts the most recent response from Codex rollout files and opens it in revdiff for review and annotation. Activates on "revdiff-plan", "review plan with revdiff", "annotate plan", "review last response"…
code-reviewer
Code review specialist focused on patterns, bugs, security, and performance.