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 vivy-yi/finance-skills --skill control-testinggit clone --depth 1 https://github.com/vivy-yi/finance-skillsWrote 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/vivy-yi/finance-skills/control-testing)<a href="https://agentmods.dev/skills/vivy-yi/finance-skills/control-testing"><img src="https://agentmods.dev/badge/skills/vivy-yi/finance-skills/control-testing/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/vivy-yi/finance-skills/control-testing"><img src="https://agentmods.dev/badge/skills/vivy-yi/finance-skills/control-testing.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 | $0.00081 | $0.02009 |
| Opus 5 | $0.00041 | $0.01005 |
| Sonnet 5 | $0.00016 | $0.00402 |
| Haiku 4.5 | $0.00008 | $0.00201 |
Grade A, and why
control-testing 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 — 219 lines — stays where its author put it; the contents beside it link to each section on GitHub.
加载上下文
首次使用时: 读取 ../../CLAUDE.md 获取场景级配置(测试标准/样本量公式/偏差容忍度)。
/control-testing — 控制测试
Examples
→ 示例:审计师说"需要对我们新设计的采购审批控制做个穿行测试,记录样本量和偏差",系统应调用本技能,执行抽样测试并输出测试记录。
→ 示例:用户说"CFO 要求在年度 SOX 审计前对所有关键控制做一次中期测试",系统应调用本技能,按控制清单执行全量测试计划。
→ 示例:用户说"系统刚上线了新 CRM,控制变了,帮我验证新控制是否有效",系统应调用本技能,执行变更后的控制测试。
测试类型定义
测试类型:
□ 询问(Inquiry):向控制执行人了解控制执行情况
□ 观察(Observation):现场观察控制执行过程
□ 检查(Inspection):审查控制执行的证据文件
□ 重新执行(Re-performance):审计人员独立执行控制并比对结果
测试方法:
□ 抽样测试:从总体中抽取样本逐一验证
□ 全量测试:对全部交易/记录执行验证(适用于小样本总体)
□ 穿行测试(Walkthrough):从头到尾完整走一遍流程验证控制
第一步:确定测试范围和样本量
测试总体确认:
□ 控制编号:[C-01]
□ 控制名称:[名称]
□ 控制频率:[每日/每周/每月]
□ 测试期间:[YYYY-MM-DD 至 YYYY-MM-DD]
□ 测试总体规模:
→ 总体笔数:[X] 笔
→ 总体金额:[X] 万
□ 样本量计算(统计抽样):
→ 预期偏差率:[X]%(历史数据)
→ 可容忍偏差率:[X]%(审计标准)
→ 置信水平:[X]%(通常 95%)
→ 最小样本量:[X] 笔
样本选取:
□ 抽样方法:[随机抽样/分层抽样/判断抽样]
□ 选取笔数:[X] 笔
□ 金额覆盖:[X]%(样本金额 / 总体金额)
第二步:执行抽样测试
对每笔样本执行检查:
□ 控制编号:[C-01] — [控制名称]
样本 #1:
□ 交易日期:[YYYY-MM-DD]
□ 交易金额:[X] 万
□ 交易描述:[内容]
□ 控制执行时间:[YYYY-MM-DD]
测试结果:
□ 证据存在:[✅ 有 / 🔴 无]
□ 证据完整性:[✅ 完整 / ⚠️ 缺失部分:描述]
□ 审批人正确:[✅ 是 / 🔴 否 — 实际审批人:姓名]
□ 审批时间合规:[✅ 是(事后 [X] 天内)/ 🔴 否]
□ 执行质量:[✅ 符合要求 / 🔴 不符合 — 偏差:描述]
测试结论:[✅ 通过 / 🔴 未通过]
第三步:测试结果汇总
样本测试明细:
| 样本# | 日期 | 金额 | 审批人 | 证据 | 时间合规 | 结论 |
|-------|------|------|--------|------|---------|------|
| 1 | [日期] | [X]万 | [姓名] | [✅/🔴] | [✅/🔴] | [✅/🔴] |
| 2 | [日期] | [X]万 | [姓名] | [✅/🔴] | [✅/🔴] | [✅/🔴] |
偏差分析:
□ 测试样本数:[X] 笔
□ 发现偏差数:[X] 笔
□ 偏差率:[X]%(偏差数 / 样本数)
□ 偏差详情:
| 样本# | 偏差描述 | 偏差类型 | 影响金额 | 是否为实质性偏差 |
|-------|---------|---------|---------|----------------|
| [#] | [描述] | [执行/设计] | [X]万 | [是/否] |
□ 偏差根本原因分析:
→ 人员能力问题:[是/否] — 描述
→ 系统/流程问题:[是/否] — 描述
→ 管理层凌驾:[是/否] — 描述
□ 偏差是否具有系统性:[是(多个样本同类偏差)/否]
第四步:测试结论
控制有效性判断:
□ 可容忍偏差率:[X]%
□ 实际偏差率:[X]%
□ 偏差率 vs 可容忍偏差率:[✅ ≤ / 🔴 >]
□ 结论判断:
→ 偏差率 ≤ 可容忍偏差率:[✅ 控制有效(In Effective)]
→ 偏差率 > 可容忍偏差率:[🔴 控制无效(In Effective,存在缺陷)]
→ 样本偏差具有系统性:[🔴 控制无效(设计缺陷或执行不一致)]
□ 特别说明(如适用):
→ 样本量不足:[是/否] — 说明
→ 发现管理层凌驾:[是/否] — [须升级]
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 · 219 lines · 81 tokens per session scan A d99aafddd024
control-testing is a skill published in the GitHub repository vivy-yi/finance-skills (27 stars, last pushed 2mo ago), licensed MIT. It adds 81 tokens to every session and 2,009 once invoked, about $0.0004 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
tdd
A Test-Driven Development guide. TDD means writing a failing test first, adding the smallest implementation that passes it, and then improving the code while keeping the tests passing.
debugging
A structured method for finding and fixing software bugs. It starts by writing a test that reproduces the failure, then investigates its underlying cause before making a small fix.
writing-skills
Use when creating new skills, editing existing skills, or verifying skills work before deployment.
solo-build
This skill is self-contained — follow the task loop, TDD rules, and completion flow below instead of delegating to external build/execution skills (superpowers, etc.).
backend-development
Backend API design, database architecture, microservices patterns, and test-driven development. Use for designing APIs, database schemas, or backend system architecture.
tdd-workflow
Test-Driven Development workflow — red-green-refactor cycle.