Borrowing it
Nothing to install: this file belongs to in-the-loop-labs/pair-review. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/in-the-loop-labs/pair-review/main/.pi/skills/review-roulette/SKILL.mdgit clone --depth 1 https://github.com/in-the-loop-labs/pair-reviewWrote 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/in-the-loop-labs/pair-review/review-roulette)<a href="https://agentmods.dev/skills/in-the-loop-labs/pair-review/review-roulette"><img src="https://agentmods.dev/badge/skills/in-the-loop-labs/pair-review/review-roulette/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/in-the-loop-labs/pair-review/review-roulette"><img src="https://agentmods.dev/badge/skills/in-the-loop-labs/pair-review/review-roulette.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00030 | $0.01218 |
| Opus 5.5 | $0.00012 | $0.00487 |
| Sonnet 5.5 | $0.00006 | $0.00244 |
| Haiku 4.5 | $0.00003 | $0.00122 |
Grade A, and why
review-roulette 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 2mo 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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review Roulette
When this skill is active, your ONLY job is orchestration — you do NOT perform any review analysis yourself. You randomly select 3 reasoning models, dispatch the review to all of them in parallel, and merge the results.
Step 1: Discover Available Reasoning Models
Run ${PI_CMD:-pi} --list-models via bash to get the current list of models with
valid API keys. Eligible models are those that show thinking: yes in the
output — these are the reasoning-capable / premium models.
Excluded models: Never select openai/o3-pro — it is prohibitively
expensive. If it appears in the model list, skip it entirely.
Example reasoning models you might see (provider/model format):
anthropic/claude-opus-4-6anthropic/claude-sonnet-4-5(with thinking)openai/o3openai/o4-miniopenai/gpt-5-proopenai/gpt-5.2-progoogle/gemini-2.5-pro(with thinking)google/gemini-2.5-flash(with thinking)google/gemini-3.1-pro-previewxai/grok-4
The exact list depends on which API keys are configured. Always check — do not assume models are available.
Step 2: Randomly Select 3 Models
From the eligible reasoning models, pick exactly 3 at random.
CRITICAL — true randomness and diversity:
- Do NOT always pick the same 3 models. The entire point of review roulette is variety of perspectives across runs.
- Prefer different providers when possible. If you have reasoning models from Anthropic, OpenAI, Google, and xAI, pick from 3 different providers. Only double up on a provider if fewer than 3 providers have eligible models.
- Shuffle or randomize your selection each time. Do not default to alphabetical order or any fixed preference.
Step 3: Dispatch the Review in Parallel
Use the task tool with the tasks array to dispatch all 3 reviews
simultaneously. Each task object must include:
model: The selected model inprovider/modelformat.task: The FULL original review prompt/instructions. Each subtask starts fresh with NO conversation history and NO context from the parent. You must forward EVERYTHING you were asked to do — the complete prompt, all instructions, the diff, file contents, any constraints or formatting requirements, the expected JSON output schema, etc. Do not summarize or abbreviate. Pass it all through verbatim.
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.
- 2mo ago First seen · 145 lines · 30 tokens per session scan A 5b6ec160d9be
review-roulette is a skill published in the GitHub repository in-the-loop-labs/pair-review (73 stars, last pushed 8d ago), licensed Apache-2.0. It adds 30 tokens to every session and 1,218 once invoked, about $0.0001 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 2026-08-30.
Other skills, from other repositories
autoreview
Structured code review through the shared OpenClaw agent-skills installation.
omh-code-review
This is an OMH code-review workflow skill, projected for Agent Skills hosts (Claude Code, Codex, Cursor, opencode, OpenClaw, pi).
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"…
explore-unknowns
Guide the user through a quadrant walk that maps the unknowns of a task — open by listing the known knowns, then work through known unknowns, unknown knowns, and unknown unknowns one stage at a time, ending with a complete four-quadrant map in the user's hands. Use when a request is ambiguous or underspecified, the…
re0-merge
Review and land an external contribution the way this suite does: gate it against the thesis, land it with the author's credit intact, complete a new skill rather than merging it raw, then approve, credit, and explain before closing. Use when reviewing a pull request, as any collaborator or maintainer, not only the…
asplos-writing-style
Use when drafting or revising ASPLOS prose — engineering the first two pages that rapid review actually reads, stating the cross-layer insight as one quotable sentence, structuring per-layer mechanism sections, writing bounded claims, and fitting everything into 11 self-contained pages that include figures and…