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 flanliulf/SpecLite --skill speclite-code-review-02-evaluatorgit clone --depth 1 https://github.com/flanliulf/SpecLiteWrote 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/flanliulf/speclite/speclite-code-review-02-evaluator)<a href="https://agentmods.dev/skills/flanliulf/speclite/speclite-code-review-02-evaluator"><img src="https://agentmods.dev/badge/skills/flanliulf/speclite/speclite-code-review-02-evaluator/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/flanliulf/speclite/speclite-code-review-02-evaluator"><img src="https://agentmods.dev/badge/skills/flanliulf/speclite/speclite-code-review-02-evaluator.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 | $0.00065 | $0.01432 |
| Opus 5 | $0.00032 | $0.00716 |
| Sonnet 5 | $0.00013 | $0.00286 |
| Haiku 4.5 | $0.00006 | $0.00143 |
Grade A, and why
speclite-code-review-02-evaluator 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 7d 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
[技能说明] 对跨 LLM 代码审查的结果进行独立评估,判断审查发现的合理性和准确性,生成结构化的评估文档。作为审查工作流的质量把关环节,确保 CR 发现的客观性。
[核心能力] - 审查结果评估:对 CR 代码审查的发现逐条评估其合理性和准确性 - 四桶分类感知:识别并利用审查结果中的「来源」(blind/edge/auditor)和「分类」(decision_needed/patch/defer)增强字段,辅助评估判断 - 自动定位最新结果:自动扫描并定位最新一轮的《代码审查结果文件》 - 评估轮次管理:自动检测已有评估轮次,正确编号新一轮评估 - 历史参考:在有明确异议时可参考过往轮次的审查结果 - 结构化评估输出:生成规范化的评估文档,包含逐条评估结论 - 只读安全保障:严格禁止修改源码、Story 文档和执行修复操作
[执行流程]
路径约定和文件名格式以 references/cr-config.md 为准。
Step 1:定位代码审查目录和文件
- 读取 `references/cr-config.md` 获取路径约定
- 按配置中的 Story ID 规则提取 `{story-id}`
- 按配置中的代码审查目录格式确定路径
- 生成数据:story-id、code-review-dir
Step 2:定位最新一轮审查结果
- 按配置中的审查总结文件名格式,扫描 code-review-dir 下匹配的文件
- 找到 round 值(n)最大的文件作为本次评估对象
- 读取该文件的完整内容
- 生成数据:latest-review-file、review-round-number
Step 3:检测评估轮次
- 按配置中的审查评估文件名格式,扫描 code-review-dir 下匹配的文件
- 统计已有的评估轮次数量,确定本轮评估轮次号 m = 已有评估轮次 + 1
- 生成数据:evaluation-round-number(m)
Step 4:执行评估
- 逐条审阅 CR 审查结果中的发现(Findings)
- 若发现包含「来源」字段(如 blind/edge/auditor/blind+edge),将其作为评估参考:
- 多来源命中(如 blind+edge)的发现可信度更高,应优先确认
- 单来源发现需更谨慎验证,尤其关注是否为误报
- 若发现包含「分类」字段(如 decision_needed/patch/defer),将其作为评估参考:
- `decision_needed`:重点评估是否确实需要人工裁决
- `patch`:验证修复方案的可行性和完整性
- `defer`:确认是否确实为既有问题而非本次改动引入
- 对每条发现进行评估:
- 问题描述是否准确?
- 严重性判断是否合理?
- 修复建议是否可行?
- 是否存在误报(false positive)?
- 如有明确异议,可参考过往轮次的《代码审查结果文件》进行交叉验证
- 给出整体评估结论:
- 哪些发现需要修复(分优先级)
- 哪些发现可以忽略(说明理由)
- 哪些发现需要进一步讨论
- 生成数据:evaluation-findings(评估结论列表)
Step 5:保存评估结果
- 读取输出格式模板:`assets/output-template.md`
- 按照模板中定义的章节结构生成评估文档
- 确定今天日期,格式为 YYYYMMDD
- 按 `references/cr-config.md` 中的文件名格式保存文件
- 严格按照模板中定义的章节结构和格式规范输出,包括:
- YAML 元信息头部(含 Review Source 和 Review Model)
- 评估总结
- 上轮问题回顾确认(被评估审查为复审时)
- 逐条发现评估(审查原文、评估结论、评估分析)
- 整体评估结论(需修复表格、CR TODO 表格、评估决定)
- 向用户展示评估结论要点
- 完成后返回:"✅ CR 代码审查结果评估完成(第 m 轮),结果已保存"
[注意事项]
- 绝对禁止修改任何源码文件
- 绝对禁止修改 Story 文档内容
- 绝对禁止自行执行修复操作
- 只对最新一轮(n 值最大)的《代码审查结果文件》进行评估
- 只有在有明确异议时才允许参考过往轮次的《代码审查结果文件》
- 路径约定和文件名格式以 references/cr-config.md 为准,不硬编码
- 始终使用中文输出评估结果
- 评估要客观公正,对误报要明确标注并给出理由
- 审查结果中的「来源」和「分类」字段为可选增强信息,若存在则利用其辅助评估,若不存在则按原有逻辑评估
- 输出文件头部的 Model Used 字段必须如实填写当前执行评估的模型名称;Review Model 字段从被评估的审查结果文件头部读取,便于跨 LLM 追溯
- 如果找不到审查结果文件,立即停止并告知用户
What ships with it
3 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.
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.
- 7d ago First seen · 90 lines · 65 tokens per session scan A 10b4d99aadc0
speclite-code-review-02-evaluator is a skill published in the GitHub repository flanliulf/SpecLite (4 stars, last pushed 2mo ago), licensed MIT. It adds 65 tokens to every session and 1,432 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 2026-09-03.
Other skills, from other repositories
Security Audit Checklist
A PASS/FAIL/N/A checklist covering injection, authorization, secrets handling, dependency risk, and input validation for reviewing a diff or codebase.
PR Standards
Pull request description, review depth, and merge standards so every PR carries enough context to review and merge safely.
eng
Platform-agnostic engineering agent with three modes: --plan (propose file changes for human approval AND write the per-feature todo tickets in the same pass), --build (write code from the todo tickets — the single and final build spec), --review (one adversarial whole-change review of the working diff, run by a…
Security Checklist
A structured PASS / FAIL / N/A checklist covering common risk categories for reviewing a diff before merge.
Agent Instruction Drift Check
Procedure for comparing an agent-facing instruction doc (AGENTS.md/CLAUDE.md-style) against the actual current code and configuration to find and flag mismatches.
Code Standards
Naming, structure, and consistency conventions for keeping a codebase readable and predictable as it grows.