Borrowing it
Nothing to install: this file belongs to Miguok/fable-harness. 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/Miguok/fable-harness/main/.claude/skills/adversarial-review/SKILL.mdgit clone --depth 1 https://github.com/Miguok/fable-harnessWrote 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/miguok/fable-harness/adversarial-review)<a href="https://agentmods.dev/skills/miguok/fable-harness/adversarial-review"><img src="https://agentmods.dev/badge/skills/miguok/fable-harness/adversarial-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/miguok/fable-harness/adversarial-review"><img src="https://agentmods.dev/badge/skills/miguok/fable-harness/adversarial-review.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 | $0.00142 | $0.01127 |
| Opus 5.5 | $0.00057 | $0.00451 |
| Sonnet 5.5 | $0.00028 | $0.00225 |
| Haiku 4.5 | $0.00014 | $0.00113 |
Grade A, and why
adversarial-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 4d 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.
What it actually says
多方抗辯流程（adversarial-review）
目的
防止「聽起來合理但其實是錯的」結論被採信。單一模型的自我檢查會系統性偏袒自己的結論，所以用三個獨立、不同鏡頭、預設立場為推翻的子代理交叉審。
執行步驟
-
整理待審包：把待審結論寫成一段自足的陳述，包含：
- 結論本身（一句話）
- 依據的證據（file:line、測試輸出、量測值）
- 影響範圍（改了什麼、誰依賴它）
- 待審對象若是多條獨立發現（如審查報告的 N 條 findings）：逐條各自抗辯、不得打包成一個總結論——打包會稀釋每條的審查解析度。
-
同一則訊息平行派出三個反方（必須用 Agent 工具、subagent_type 分別為
skeptic、red-team、simplifier，一次三發不得串行）。 委派抗辯三反方（skeptic/red-team/simplifier）一律明指 model: opus；唯一例外：主迴圈模型 ID 含 opus 或 fable 時可不指定 model（繼承主迴圈）；無法判斷主迴圈是誰→用 opus。（抗辯裁決由主迴圈自做，不受此限。）每個 agent 的 prompt = 待審包原文 + 該鏡頭的任務指示。逐條抗辯時每條發現固定耗 3 個子代理（三鏡頭），單輪平行子代理總數上限 6＝一輪最多 2 條發現；更多條就分批送審，批與批的結論分開回報、不合併。 反方要實跑 hook／閘門探針時，prompt 裡固定附一句：探針一律驅動複製到 scratchpad 的副本，不得直接餵.claude/hooks/的生產檔——生產檔的屍檢寫在它自己旁邊的.gate_fail，探針的合成輸入會把雜訊寫進去（2026-10-25 清出的 6 筆全是探針）。 -
裁決（過半存活制）：
- 3 票 SURVIVED → confirmed，可直接採信
- 2 票 SURVIVED → confirmed，但必須把那 1 票 REFUTED 的理由列入風險清單回報用戶
- ≤1 票 SURVIVED → 結論擋回，依 REFUTED 理由修正後重新送審
-
Loop-until-dry 與規模校準：影響生產/全域佈署/資料安全的重大結論——修正後重審，直到連續 2 輪無新 REFUTED 理由才收工；其餘結論一輪三鏡頭即收工。無法明確排除上述影響者，視同重大。每輪把「已審過的理由」附進 prompt 避免重複發現。
-
回報格式（給用戶的最終訊息必含）：
鏡頭 verdict 關鍵理由 skeptic ... ... red-team ... ... simplifier ... ... 加一行結論：
抗辯結果：N/3 存活 → confirmed / 擋回（原因）。
禁止事項
- 不得因為趕時間跳過任何一個鏡頭。
- 不得把三個鏡頭合併成一個 agent 跑（獨立性是抗辯的前提）。
- REFUTED 理由不得默默吞掉，必須回報或修正。
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.
- 4d ago Changed · +1 lines 8bf19e962cbb
- 2mo ago First seen · 43 lines · 142 tokens per session scan A 8af16c073453
adversarial-review is a skill published in the GitHub repository Miguok/fable-harness (205 stars, last pushed 4d ago), licensed MIT. It adds 142 tokens to every session and 1,127 once invoked, about $0.0006 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…