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/backtocimacoppi/praxis/codex-reviewnpx skills add BackToCimaCoppi/Praxis --skill codex-reviewgit clone --depth 1 https://github.com/BackToCimaCoppi/PraxisWrote 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/backtocimacoppi/praxis/codex-review)<a href="https://agentmods.dev/skills/backtocimacoppi/praxis/codex-review"><img src="https://agentmods.dev/badge/skills/backtocimacoppi/praxis/codex-review.svg" alt="Measured on agentmods" 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 | $0.00067 | $0.10080 |
| Opus 5 | $0.00034 | $0.05040 |
| Sonnet 5 | $0.00013 | $0.02016 |
| Haiku 4.5 | $0.00007 | $0.01008 |
Grade A, and why
codex-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 6d 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 — 435 lines — stays where its author put it; the contents beside it link to each section on GitHub.
代码评审(codex-review)
本 skill 是施工链路的收尾核验层,用户级通用,可跨项目使用。
它解决的问题:施工完成后需要一道独立检查,但这道检查最危险的失败模式不是"漏看 bug",而是"把已拍板的正确做法当成 bug 报上来、又被照着改掉"——评审者缺乏前面多轮讨论的背景,很容易把刻意的取舍误判为缺陷。而最隐蔽的一种是:评审者根本不知道某做法是已拍板决策,于是理直气壮把它标成"缺陷"。 本 skill 的全部设计都围绕堵住这个口子,核心是一条默认方向:不改除非已证实是缺陷——证据责任压在"改"这一侧,不是压在"不改"这一侧。
作用 = 照答案做低成本符合性核验:标准答案是冻结的正式 L1–L7 + 规格物化覆盖报告 + goal 章程 + 工程红线。轻量设计只解释
SD/DP的由来,不能覆盖正式规格。若两者不一致,说明 T5 物化失效,应返闸,不由评审者选边。
在施工链路里的位置:
轻量设计(SD-x) → 正式 L7(AC-x) → 真值收敛与规格冻结(覆盖报告全绿)
↓
goal 章程(红队评审过) → goal 自主执行:代码 + 测试 + 交付
↓
【本 skill】codex 单模型符合性核验 ← 用户手动触发(goal 之外)
↓
分流责任人按"不改除非已证实"分流 ── A 改 / B 待裁 / C 上浮
↓
触承重面 / 上游规格错了?──是──→ 回炉(退回设计步,见 §6)
否
↓
修完收口
不是门禁。 它不卡开工、不卡交付;只在用户说"跑一遍代码评审"时执行。
它在 goal 之外,这是有意的。 goal 内部的审核只能是 loop 闭环内的 AI 自审 + 机器检查器——引入 loop 外的观点会带来场景不同、上下文不同的噪音。本 skill 是事后的、用户主动要的第二双眼睛,不是 goal 的工序。
0. 与 adversarial-review 的分工(单模型够用的边界)
| 维度 | adversarial-review(已有) | codex-review(本 skill) |
|---|---|---|
| 评审什么 | 设计 / 用例 / goal 章程 / 任意方案,或高风险代码 | 已写完的代码 diff(低中风险符合性核验) |
| 时机 | 开工前(真值待定,找最优) | 开工后(真值已定,查偏差) |
| 真值状态 | 待定——正在探索方案空间 | 已冻结——规格 + 决策表 + 用例已拍板 |
| 模型编制 | 默认 Fable 5 + GPT-5.6-Sol 双席对抗(可点选更多) | GPT-5.5 一席,无对抗 |
| 收口者 | 当前主线程直接裁决(不设裁判) | 分流责任人(对照冻结真值分流) |
| 产出 | 采纳/驳回裁决报告 | 三级分类意见(不自动改码) |
单模型在"符合性核验"这件事上够用,但不要把它当充分代码审查:
符合性核验有标准答案(冻结规格 + 决策表 + 用例规格 + 红线就是答案),是对照答案查偏差,一个强模型够了;硬上多方对抗既贵又会把"已拍板的取舍"翻出来重吵——那正是本 skill 要堵的风险。
但代码评审里最高价值的发现常常是"答案没覆盖到的地方出了 bug"——并发竞态、边界态、错误处理、性能。这些规格通常不下沉到逻辑级(实现细节属代码自由范围),需要正交视角(adversarial-review 的强项),单模型会系统性漏检。所以本 skill 定位是低成本符合性核验,不宣称"一个强模型就够全部代码审查";这类风险按 §1 升档。
单模型也没有裁判席兜 GPT 的误报——这个位置由分流责任人的对照真值分流(§5)补上,而分流的默认方向是"不改除非已证实",证据责任在"改"侧。
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.
- 6d ago First seen · 435 lines · 67 tokens per session scan A 72c4776b400e
codex-review is a skill published in the GitHub repository BackToCimaCoppi/Praxis (6 stars, last pushed 12d ago), licensed Apache-2.0. It adds 67 tokens to every session and 10,080 once invoked, about $0.0003 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
prompt-master
Generates optimized prompts for AI tools. Activates only when the user explicitly asks to write, fix, improve, or adapt a prompt for a specific AI tool (LLM, Cursor, Midjourney, image AI, video AI, coding agents, etc.). Does not activate for general conversation, coding tasks, document writing, or other…
prd-taskmaster
Zero-config goal-to-tasks engine (the Atlas engine). Takes any goal (software, pentest, business, learning), runs adaptive discovery via brainstorming, generates a validated spec, parses into TaskMaster tasks, and hands off to execution. Use when user says "PRD", "product requirements", "I want to build", invokes…
browser-tools
Security wrapper over the upstream agent-browser skill, adding URL blocklisting, rate limiting, robots.txt enforcement, and scraping guardrails. Use when automating browser workflows that need safety limits.
auto
Intent-classified router, the front door to OrchestKit and the DEFAULT entry point for any goal-shaped request. Classifies a plain-English goal and routes it to the right specialist skill. Routing is never overhead, so use it even when the target skill seems obvious; skip only when already executing inside another…
rag-retrieval
Retrieval-Augmented Generation patterns for grounded LLM responses. Use when building RAG pipelines, embedding documents, implementing hybrid search, contextual retrieval, HyDE, agentic RAG, multimodal RAG, query decomposition, reranking, or pgvector search.
ci-debug
Diagnose a failing CI run against an 11-pattern playbook. Classifies the failure, cites the relevant memory entry, proposes the exact fix command — but NEVER applies without explicit user approval. Use when a specific PR check or GitHub Actions run failed and you want a diagnosis instead of speculation. Don't use for…