Not in the top 1,000 by GitHub stars
RickyTong1/audit-harness
Three-layer audit enforcement framework for AI agents — hooks, skills, context recovery, and audit-driven daily reports
History
Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.
View weekly changes
skills/audit-end/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
89406743ac7f → 89406743ac7f
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 89406743ac7f
No earlier version was observed for comparison.
skills/audit-init/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
89406743ac7f → 89406743ac7f
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 89406743ac7f
No earlier version was observed for comparison.
skills/audit-recover/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
89406743ac7f → 89406743ac7f
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 89406743ac7f
No earlier version was observed for comparison.
skills/audit-report-daily/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
89406743ac7f → 89406743ac7f
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 89406743ac7f
No earlier version was observed for comparison.
skills/audit-start/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
89406743ac7f → 89406743ac7f
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 89406743ac7f
No earlier version was observed for comparison.
Verified commit 89406743ac7f0464b32082a4a8ff9c95d0a6798a
git clone --filter=blob:none --no-checkout https://github.com/RickyTong1/audit-harness.git .hotskills/rickytong1-audit-harness && git -C .hotskills/rickytong1-audit-harness checkout --detach 89406743ac7f0464b32082a4a8ff9c95d0a6798a
# Copy each selected skill directory from .hotskills/rickytong1-audit-harness into .claude/skills/<skill-name>/Heat
0.3- Star growth, last 7 daysNot available
- Star growth, last 24 hoursNot available
- Rank movement, last 7 daysNot available
- Commits, last 30 daysNot available
- Accounts asking about it through MCP, last 7 daysNot available
- Hot votes this weekNot available
Heat never changes the order by stars, and sponsorship never touches it. How heat works
Stars over time
30-day change not yet available: not enough history.
Individual skills in this repository
- audit-end
skills/audit-end - audit-init
skills/audit-init - audit-recover
skills/audit-recover - audit-report-daily
skills/audit-report-daily - audit-start
skills/audit-start
How it works
audit-end
When your agent uses it
结束审计会话，汇总审计数据，执行完整性检查，保存到 .claude/runs/。当用户说"完成了"、"结束"、"暂停"、"今天到这"，或当前任务已完成需要保存审计记录时使用。
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,252 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 触发方式
- 执行步骤
Files
../../.claude-plugin/marketplace.jsonasset1,060 bytesSKILL.mdcore4,264 bytesSKILL.mdcore4,264 bytes
Read SKILL.md
# /end — 审计会话结束 + 完整性检查
## 触发方式
```
/end
```
## 执行步骤
### 1. 确定当前 session
读取 `.claude/runs/.current_session` 获取 session_id。
audit-init
When your agent uses it
审计框架安装向导。在新项目中首次使用 audit-harness 时调用。自动完成文件复制、配置生成、CLAUDE.md 注入、目录创建，开箱即用。
What it loads into context
Frontmatter
90 tokensAlways available to the agent
SKILL.md
2,123 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 触发方式
- 概述
- 执行步骤
- 幂等性
- 注意事项
Files
SKILL.mdcore7,510 bytesSKILL.mdcore7,510 bytesSKILL.mdcore7,510 bytesSKILL.mdcore7,510 bytesSKILL.mdcore7,510 bytesSKILL.mdcore7,510 bytes
Read SKILL.md
# /audit-init — 审计框架安装向导
## 触发方式
```
/audit-init
```
## 概述
本 Skill 是一个交互式安装向导（wizard），引导用户在当前项目中完成 audit-harness 的全部配置。安装完成后，`/start`、`/end`、`/recover`、`/report-daily` 四个 Skill 立即可用。
## 执行步骤
audit-recover
When your agent uses it
从审计记录恢复丢失的 Context。当你无法回忆前期讨论内容、不确定当前任务进度、要提出可能已被用户否定的方案、或用户说"你之前说的是什么"、"回忆一下"、"我们之前讨论了什么"时使用。更重要的是：当你感觉自己的回答变得笼统、缺少具体数字和文件名时，应主动触发此 Skill，而不是凭印象继续工作。
What it loads into context
Frontmatter
131 tokensAlways available to the agent
SKILL.md
952 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 触发方式
- 核心原则
- 何时触发
- 执行步骤
- 注意事项
Files
SKILL.mdcore3,477 bytes
Read SKILL.md
# /recover — Context 恢复
## 触发方式
```
/recover # 恢复当前 session 的 context
/recover full # 恢复当前 session + 最近 3 天的完整 context
/recover {session_id} # 恢复指定 session 的 context
```
## 核心原则
**不确定就查审计记录，不要凭印象工作。**
audit-report-daily
When your agent uses it
基于 .claude/runs/ 目录的审计数据自动生成工作日报。日报每个数字都可追溯到审计记录。当用户说"生成日报"、"今天做了什么"、"写报告"、"总结一下今天的工作"，或通过 cron 每日 08:03 自动触发时使用。
What it loads into context
Frontmatter
113 tokensAlways available to the agent
SKILL.md
1,192 tokensWhen the skill is activated
Reference documents
803 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 触发方式
- 数据源
- 执行步骤
- 注意事项
Files
SKILL.mdcore3,922 bytesSKILL.mdcore3,922 bytesSKILL.mdcore3,922 bytesreferences/report_template.mddoc2,482 bytes
Read SKILL.md
# /report-daily — 审计驱动工作日报
## 触发方式
```
/report-daily # 生成今天的日报（截止到当前时刻）
/report-daily 20260319 # 生成指定日期的日报
/report-daily review # 生成日报 + 晨间自我修正报告
```
## 数据源
日报的**每个数字**都有明确的数据源，不允许"凭记忆"填写。
audit-start
When your agent uses it
创建审计会话并自动恢复历史上下文。当用户开始新任务、说"帮我做X"、"分析Y"、"修改Z"、"开始工作"等，或当你检测到当前没有活跃的审计会话时，主动使用此 Skill。即使用户没有显式说 /start，只要涉及数据处理、代码修改、分析决策、文档编写，都应建议或自动启动审计会话。
What it loads into context
Frontmatter
136 tokensAlways available to the agent
SKILL.md
1,081 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 触发方式
- 为什么需要 /start
- 执行步骤
Files
SKILL.mdcore3,803 bytesSKILL.mdcore3,803 bytes
Read SKILL.md
# /start — 审计会话启动 + Context 恢复
## 触发方式
```
/start "任务描述"
```
## 为什么需要 /start
每次工作都应该在审计会话中进行。没有会话意味着：
- hooks 产出的审计数据归入自动创建的 `auto_` session，缺少任务描述
- /end 无法做完整性检查
- 日报中的任务总览缺少描述Skills included
- audit-end
- audit-init
- audit-recover
- audit-report-daily
- audit-start
Repository facts
- License
- MIT
- Created
- Mar 25, 2026
- Last pushed
- Jun 10, 2026
- Forks
- 3
- Open issues
- 3
- Project activity
- Slowing
Before you install
Network or broad shellAutomated signals, not a guarantee
Checked Nov 6, 2026
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 9 notes
skills/audit-end/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedskills/audit-init/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedskills/audit-init/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedskills/audit-init/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedskills/audit-init/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedskills/audit-init/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedskills/audit-report-daily/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedskills/audit-report-daily/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedskills/audit-start/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- 4 broad permissionsin 4 files: audit-end/SKILL.md:1 audit-init/SKILL.md:1 audit-report-daily/SKILL.md:1 and 1 more
- 1 dangerous shell commandin 1 file: audit-end/SKILL.md:135
- No network access
- No hidden or obfuscated text
- No instruction overrides
- No executable scripts
- No access to secrets
- No requests to vote in Hot
Show all 5 signals by file
skills/audit-end/SKILL.md
skills/audit-init/SKILL.md
- Broad permissionsline 1
Bash
skills/audit-report-daily/SKILL.md
- Broad permissionsline 1
Bash
skills/audit-start/SKILL.md
- Broad permissionsline 1
Bash
How the signals fit together
Facts from a static read of the files at the pinned commit. They describe what the code does, not why.
Some files could not be read; the facts below still hold.
Declared and executed: 1 fact
- Allows Bash without restrictions, though it has no scripts or shell blocksaudit-init/SKILL.md:5High confidence
Signal confidence: 5 facts
- Broad permissions in the frontmatter of the SKILL.mdaudit-end/SKILL.md:1High confidence
- Dangerous shell operation in code of the SKILL.mdaudit-end/SKILL.md:135High confidence
- Broad permissions in the frontmatter of the SKILL.mdaudit-init/SKILL.md:1High confidence
- Broad permissions in the frontmatter of the SKILL.mdaudit-report-daily/SKILL.md:1High confidence
- Broad permissions in the frontmatter of the SKILL.mdaudit-start/SKILL.md:1High confidence
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Native | Manifest .claude-plugin/marketplace.json | .claude/skills/<skill-name>/ |
| Codex | Compatible | Standard SKILL.md skills/audit-end/SKILL.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md skills/audit-end/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md skills/audit-end/SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md skills/audit-end/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md skills/audit-end/SKILL.md | .opencode/skills/<skill-name>/ |
- Native The repository ships a manifest, folder or frontmatter field for this agent.
- Compatible A standard SKILL.md that this agent can read.
- Unknown No evidence either way.
Derived automatically from paths, manifests and frontmatter. Not tested on each agent.
