← Back to Skills Marketplace
johnsmithfan

AI Skill Maintainer (EN)

by JohnSmithfan · GitHub ↗ · v1.1.0-en2 · MIT-0
linuxdarwinwin32 ⚠ suspicious
126
Downloads
0
Stars
0
Active Installs
3
Versions
Install in OpenClaw
/install ai-skill-maintainer
Description
AI公司 Skill 维护工作流(CTO 版本govern + CISO security运营standard版)。当需要对已publish的 Skill 进行版本update、bug修复、Function增强、依赖upgrade、security补丁、废弃(deprecation)manage时trigger。...
README (SKILL.md)

\r \r

AI Skill 维护工作流(CTO × CISO standard)\r

\r

executerole:Skill 维护者(CTO 版本govern + CISO security运营)\r 版本:v1.0.0(CTO-001 版本govern × CISO-001 security运营)\r compliance状态:✅ 维护操作需record,🚨 security补丁走紧急通道\r \r ---\r \r

核心principle\r

\r

  1. 变更可trace:所有修改必须record版本历史\r
  2. 向后兼容:MINOR/PATCH 变更不得破坏现有Function\r
  3. security优先:CISO security补丁走紧急通道,不受正常publishcyclerestrict\r
  4. 最小变更:只改必要的,不要过度工程化\r \r ---\r \r

Agent 调用接口(Inter-Agent Interface)\r

\r

版本:v1.1.0(新增接口层)\r securityConstraint:接口本身零新增攻击面,所有输入参数均经过verify\r \r ---\r \r

接口身份\r

\r | 属性 | 值 |\r |------|-----|\r | 接口 ID | skill-maintainer-v1 |\r | 调用方式 | sessions_send / sessions_spawn (isolated) |\r | 会话Goal | isolated(强制隔离)|\r | 最低permission | L3(可读写 skills/ 指定目录) |\r | CISO Constraint | 🚨 security补丁任务必须 CISO-001 authorize,紧急通道优先 |\r \r ---\r \r

TASK 消息格式\r

\r

{\r
  "skill": "ai-skill-maintainer",\r
  "version": "1.1.0",\r
  "task": "\x3Ctask-type>",\r
  "params": { ... },\r
  "context": {\r
    "caller": "\x3Ccaller-agent-id>",\r
    "priority": "\x3CP0|P1|P2|P3>",\r
    "emergency": false,\r
    "isolated": true\r
  }\r
}\r
```\r
\r
### 可用 Task 类型\r
\r
| Task | 参数 | 返回 | Description |\r
|------|------|------|------|\r
| `diagnose` | `skill-name`, `issue`, `caller` | `{diagnosis, type, severity}` | 诊断问题 |\r
| `patch` | `skill-name`, `version`, `changes`, `caller` | `{new-version, status}` | implement修复 |\r
| `security-patch` | `skill-name`, `cve-id`, `authorization`, `caller` | `{fixed, new-version, notification-sent}` | 🚨 CVE 修复 |\r
| `deprecate` | `skill-name`, `reason`, `replacement`, `caller` | `{deprecated-version, status}` | 废弃 Skill |\r
| `emergency-isolate` | `skill-name`, `reason`, `caller` | `{isolated, affected-versions}` | 🚨 紧急隔离(0-day)|\r
| `health-check` | `skill-name` | `{metrics, status}` | 健康检查 |\r
| `dependency-audit` | `skill-name` | `{dependencies[], cves[]}` | 依赖 CVE 扫描 |\r
\r
### Task 参数 Schema\r
\r
#### `security-patch` 参数\r
\r
```json\r
{\r
  "skill-name":     "string (required, skill slug)",\r
  "cve-id":         "string (required, e.g. CVE-YYYY-NNNNN)",\r
  "cvss-score":     "number (required, 0.0-10.0)",\r
  "authorization":  "string (required, must be CISO-001)",\r
  "caller":         "string (required, agent ID)",\r
  "changes": {\r
    "description":  "string (what was changed)",\r
    "files":        "string[] (list of modified files)",\r
    "test-results": "string (test outcome summary)"\r
  }\r
}\r
```\r
\r
**CVE 紧急通道 SLA**:\r
\r
| CVSS | trigger | SLA | process |\r
|------|------|-----|------|\r
| 9.0-10.0 | 🚨 紧急隔离 + Patch | ≤ 24h | 紧急通道直通 |\r
| 7.0-8.9 | 紧急 Patch | ≤ 7d | standard通道加速 |\r
| 4.0-6.9 | 计划修复 | ≤ 30d | standard通道 |\r
| 0.1-3.9 | track | 下个版本 | 常规process |\r
\r
#### `emergency-isolate` 参数\r
\r
```json\r
{\r
  "skill-name": "string (required)",\r
  "reason":     "string (required, CVE ID or incident description)",\r
  "caller":     "string (required, must be CISO-001 or CTO-001)"\r
}\r
```\r
\r
#### `health-check` 参数\r
\r
```json\r
{\r
  "skill-name": "string (required)"\r
}\r
```\r
\r
> **健康threshold**(CTO-001 KPI 对齐):TSR \x3C 92% → `UNHEALTHY`;P95 > 1200ms → `DEGRADED`;CVSS ≥ 7.0 → `HIGH_RISK`\r
\r
**返回值示例**:\r
```json\r
{\r
  "status": "success",\r
  "result": {\r
    "skill-name": "pdf-processor",\r
    "metrics": {\r
      "tsr": 94.2,\r
      "p95-latency-ms": 850,\r
      "cvss-score": 3.8\r
    },\r
    "status": "HEALTHY",\r
    "recommendations": []\r
  }\r
}\r
```\r
\r
#### `dependency-audit` 参数\r
\r
```json\r
{\r
  "skill-name": "string (required)"\r
}\r
```\r
\r
**返回值示例**:\r
```json\r
{\r
  "status": "success",\r
  "result": {\r
    "dependencies": [\r
      {"name": "requests", "version": "2.28.0", "latest": "2.32.0"}\r
    ],\r
    "cves": [\r
      {"id": "CVE-2024-XXXX", "severity": "critical", "fix": "upgrade to 2.32.0+"}\r
    ]\r
  }\r
}\r
```\r
\r
**隔离决策verify**:\r
```python\r
# 伪代码\r
authorized = params["caller"] in {"CISO-001", "CTO-001"}\r
has_reason = len(params["reason"]) > 10\r
if not authorized:\r
    raise PermissionError("Only CISO-001 or CTO-001 can trigger emergency isolate")\r
if not has_reason:\r
    raise ValueError("Emergency isolate requires documented reason")\r
```\r
\r
### 返回值 Schema\r
\r
```json\r
{\r
  "status":   "success | error | pending | isolated",\r
  "task":     "\x3Ctask-type>",\r
  "result":   { ... },\r
  "meta": {\r
    "reviewer":    "\x3Cagent-id>",\r
    "duration-ms": "\x3Celapsed>",\r
    "cve-resolved": "\x3CCVE-ID if security-patch>",\r
    "sla-status":  "WITHIN_SLA | BREACHING | RESOLVED"\r
  }\r
}\r
```\r
\r
### 错误码\r
\r
| Code | Meaning | Action |\r
|------|---------|--------|\r
| `E_SKILL_NOT_FOUND` | Skill 不存在 | 返回可用版本列表 |\r
| `E_UNAUTH_PATCH` | 未authorizesecurity补丁 | reject,notify CISO |\r
| `E_CVE_SLA_BREACH` | CVE SLA 即将/已违约 | 上报 CTO + CISO |\r
| `E_ISOLATE_CONFLICT` | 已在隔离状态 | 返回当前状态 |\r
| `E_DEPENDENCY_CVE` | 依赖含已知 CVE | 返回 CVE 详情和修复建议 |\r
| `E_VERSION_CONFLICT` | 版本号冲突 | 返回正确版本号建议 |\r
\r
### Agent 间调用示例\r
\r
```markdown\r
# CTO-001 请求诊断\r
sessions_send(sessionKey="cto-isolated", message="\r
skill: ai-skill-maintainer\r
task: diagnose\r
params:\r
  skill-name: pdf-processor\r
  issue: User reports skill crashes when processing large files\r
  caller: CTO-001\r
priority: P2\r
isolated: true\r
")\r
\r
# CISO-001 请求 CVE 紧急修复\r
sessions_send(sessionKey="ciso-isolated", message="\r
skill: ai-skill-maintainer\r
task: security-patch\r
params:\r
  skill-name: pdf-processor\r
  cve-id: CVE-2026-12345\r
  cvss-score: 9.1\r
  authorization: CISO-001\r
  caller: CISO-001\r
  changes:\r
    description: Fixed command injection via path parameter\r
    files: [scripts/process.py]\r
    test-results: All regression tests pass\r
emergency: true\r
")\r
\r
# CQO-001 请求健康检查\r
sessions_send(sessionKey="cqo-isolated", message="\r
skill: ai-skill-maintainer\r
task: health-check\r
params:\r
  skill-name: pdf-processor\r
")\r
```\r
\r
### securityConstraint(接口层)\r
\r
```\r
🚨 接口security红线:\r
• skill-name 参数仅接受 [a-z0-9-] 字符,reject斜杠/点号(防止path注入)\r
• authorization 字段仅接受 CISO-001 签名的security任务\r
• emergency-isolate 仅接受 CISO-001 或 CTO-001 authorize\r
• 隔离execute:所有 agent 调用必须在 isolated 会话中运行\r
• CVE respond:CVSS ≥ 9.0 必须 15 分钟内respond,否则 SLA 违约alert\r
• 日志脱敏:返回结果不得含 caller 私人data\r
```\r
\r
### 与其他 Skill 的接口关系\r
\r
| 调用方 | Task | trigger条件 |\r
|--------|------|---------|\r
| **CTO-001** | `diagnose`, `patch`, `emergency-isolate` | 版本manage/紧急respond |\r
| **CISO-001** | `security-patch`, `emergency-isolate`, `dependency-audit` | CVE handle/security incident |\r
| **CQO-001** | `health-check`, `diagnose` | 质量monitor |\r
| **ai-skill-creator** | `patch` (子 Skill) | 创作process中discover bug |\r
| **ai-skill-optimizer** | `dependency-audit` | optimize前基线检查 |\r
\r
---\r
\r
## 维护场景分类\r
\r
| 场景 | trigger关键词 | 版本upgrade | securityreview |\r
|------|-----------|---------|---------|\r
| Bug 修复 | "修复 bug"、"修复错误" | PATCH | 正常 |\r
| Function增强 | "增强Function"、"新增Function" | MINOR | 正常 |\r
| 不兼容变更 | "Breaking Change"、"重构" | MAJOR | 正常 |\r
| 依赖security补丁 | "security补丁"、"CVE 修复" | PATCH | 🚨 紧急通道 |\r
| 废弃notify | "废弃Skill"、"停用" | PATCH | 正常 |\r
\r
---\r
\r
## standard维护process(5步)\r
\r
### Step 1 — 诊断(Diagnosis)\r
\r
**输入**:用户描述的问题或需求\r
\r
**诊断清单**:\r
\r
```markdown\r
## 诊断record\r
\r
Skill 名称:\x3Cname>\r
当前版本:\x3Cversion>\r
问题类型:[Bug / Function缺失 / security漏洞 / 依赖过时 / 其他]\r
\r
### 问题描述\r
\x3C用户描述>\r
\r
### 复现step(如适用)\r
1.\r
2.\r
3.\r
\r
### 影响范围\r
- 影响的Function:\r
- 影响的用户/Agent:\r
\r
### 初步判断\r
- 根因:\r
- 修复plan:\r
- 版本影响:[PATCH / MINOR / MAJOR]\r
```\r
\r
**CISO security场景判断**:\r
\r
| 判断条件 | 结论 | process |\r
|---------|------|------|\r
| 涉及 CVE/漏洞 | 🚨 security紧急 | 跳至security补丁process |\r
| 涉及凭证泄露 | 🚨 security紧急 | 立即notify + 紧急修复 |\r
| 涉及 PII 泄露 | 🚨 security紧急 | 立即notify + 紧急修复 |\r
| 其他 | ✅ 正常维护 | 继续standardprocess |\r
\r
---\r
\r
### Step 2 — analyze(Analysis)\r
\r
**输出**:[references/maintenance-log.md](references/maintenance-log.md) record\r
\r
#### 2.1 变更范围analyze\r
\r
```markdown\r
### 受影响文件\r
| 文件 | 变更类型 | riskassess |\r
|------|---------|---------|\r
| SKILL.md | [修改/新增/删除] | 🟢 低 |\r
| scripts/*.py | ... | ... |\r
\r
### 兼容性影响\r
- 向后兼容:✅ / ❌\r
- trigger关键词变更:✅ / ❌(如有变更需notify用户)\r
- 工具permission变更:✅ / ❌\r
\r
### 测试计划\r
- [ ] 本地测试用例:\r
- [ ] 回归测试:\r
```\r
\r
#### 2.2 security影响analyze(CTO + CISO)\r
\r
| analyze维度 | 检查项 | 结论 |\r
|---------|--------|------|\r
| **Function影响** | 修改是否改变核心Function? | |\r
| **permission影响** | permission是否变更? | |\r
| **依赖影响** | 依赖是否新增/upgrade/删除? | |\r
| **data影响** | 是否影响datahandle? | |\r
| **security影响** | 变更是否影响security边界? | |\r
\r
---\r
\r
### Step 3 — implement(Implementation)\r
\r
#### 3.1 版本号update\r
\r
```bash\r
# 根据变更类型确定版本\r
# Bug 修复          → vX.Y.Z → vX.Y.(Z+1)\r
# Function增强          → vX.Y.Z → vX.(Y+1).0\r
# Breaking Change   → vX.Y.Z → (X+1).0.0\r
# security补丁          → vX.Y.Z → vX.Y.(Z+1)  (强制)\r
```\r
\r
#### 3.2 SKILL.md update\r
\r
**update Frontmatter 版本**:\r
```yaml\r
---\r
name: \x3Cskill-name>\r
version: X.Y.Z   # ← update版本号\r
description: |   # ← 如有变更同步update\r
  ...\r
---\r
```\r
\r
**update版本历史**(在文件顶部或底部):\r
```markdown\r
## 版本历史\r
\r
| 版本 | 日期 | Changes |\r
|------|------|---------|\r
| X.Y.Z | YYYY-MM-DD | \x3C变更摘要> |\r
| ... | ... | ... |\r
```\r
\r
#### 3.3 scripts/ update\r
\r
**update检查清单**:\r
```markdown\r
- [ ] 脚本已update\r
- [ ] 脚本版本号已update(如有版本mechanism)\r
- [ ] 依赖已update(如有)\r
- [ ] 新增依赖已record\r
- [ ] 脚本测试已通过\r
```\r
\r
---\r
\r
### Step 4 — securityreview(Security Review)\r
\r
> ⚠️ **强制门禁**:所有变更必须通过 CISO securityreview\r
\r
#### 4.1 变更 diff review\r
\r
**reviewChanges**(对比上1版本):\r
- 新增的代码是否含 RED FLAGS?\r
- 修改的代码是否引入新漏洞?\r
- 删除的代码是否影响security边界?\r
\r
#### 4.2 依赖review\r
\r
**检查依赖变更**:\r
```bash\r
# 列出新增/upgrade的依赖\r
# 检查 CVE\r
```\r
\r
**CVE respond SLA**:\r
\r
| CVSS | 严重性 | 修复 SLA |\r
|------|--------|---------|\r
| 9.0-10.0 | Critical | 24小时 |\r
| 7.0-8.9 | High | 7天 |\r
| 4.0-6.9 | Medium | 30天 |\r
| 0.1-3.9 | Low | 下个版本 |\r
\r
#### 4.3 security补丁紧急通道\r
\r
**trigger条件**:discover Critical/High CVE\r
\r
```\r
🚀 紧急通道process:\r
\r
1. 立即隔离:停止问题版本分发\r
2. assess影响:确定受影响的 Skill 和版本\r
3. 紧急修复:最短path修复漏洞\r
4. 快速review:CISO 紧急review(可跳过部分正常process)\r
5. 紧急publish:Patch 版本,立即publish\r
6. 用户notify:notify所有受影响用户\r
```\r
\r
---\r
\r
### Step 5 — verify与publish(Verify & Publish)\r
\r
#### 5.1 verify清单\r
\r
```markdown\r
## publish前verify\r
\r
- [ ] Changes与诊断1致\r
- [ ] 版本号符合变更类型\r
- [ ] securityreview通过\r
- [ ] 脚本测试通过\r
- [ ] changelog 已update\r
- [ ] SKILL.md 已同步update\r
```\r
\r
#### 5.2 publish命令\r
\r
```bash\r
# 打包\r
clawhub package ./\x3Cskill-name> --output ./dist\r
\r
# publish\r
clawhub publish ./\x3Cskill-name> \\r
  --slug \x3Cskill-name> \\r
  --name "\x3CSkill Name>" \\r
  --version X.Y.Z \\r
  --changelog "\x3C变更摘要>"\r
```\r
\r
#### 5.3 notify(如有必要)\r
\r
```markdown\r
## 用户notify\r
\r
如有 Breaking Change 或重要security修复:\r
- notify方式:在 Skill 描述中注明\r
- notify内容:\r
  • 变更摘要\r
  • upgrade建议\r
  • 兼容性问题(如有)\r
```\r
\r
---\r
\r
## 版本历史(Changelog)\r
\r
| 版本 | 日期 | Changes | 审核人 |\r
|------|------|---------|--------|\r
| **1.1.0** | 2026-04-13 | 新增 Agent 调用接口层(Inter-Agent Interface):7个 Task 类型(diagnose/patch/security-patch/deprecate/emergency-isolate/health-check/dependency-audit);CVE 紧急通道 SLA system;emergency-isolate authorizeverify;与 ai-skill-creator / ai-skill-optimizer 接口关系Definition | CTO-001 / CISO-001 |\r
| **1.0.0** | 2026-04-11 | Initial version:CTO 版本govern5步维护process + CISO security运营standard(漏洞respond SLA + 补丁manage)+ 废弃manageprocess | CTO-001 / CISO-001 |\r
\r
## rollbackstrategy(Rollback)\r
\r
> 如维护操作失败,execute以下steprecover:\r
\r
```bash\r
# recover到上1个可用版本\r
git checkout tags/v\x3C上1版本> -- SKILL.md scripts/ references/\r
\r
# verifyrollback成功\r
git log --oneline -3\r
```\r
\r
**rollbacktrigger条件**:\r
- `emergency-isolate` 后:满足 CVE 已修复 + CISO-001 复审通过 + CQO-001 验收通过后方可解除隔离\r
- `patch` 失败:rollback到隔离前版本,notify CTO-001\r
- `deprecate` 误操作:recover `deprecated: false`,notify CRO-001\r
\r
---\r
\r
## 废弃(Deprecation)manage\r
\r
### 废弃process\r
\r
```\r
废弃notify(vX.Y.Z)→ 过渡期(建议 30天)→ 正式废弃(vX.Y.Z+1)\r
```\r
\r
### 废弃 SKILL.md 模板\r
\r
```markdown\r
---\r
name: \x3Cdeprecated-skill>\r
version: X.Y.Z\r
description: |\r
  ⚠️ 【已废弃】此 Skill 已废弃,建议使用 `\x3Cnew-skill>`。\r
  废弃日期:YYYY-MM-DD\r
  最后支持日期:YYYY-MM-DD\r
  迁移指南:见 references/migration.md\r
deprecated: true\r
replacement: \x3Cnew-skill-name>\r
metadata:\r
  {"openclaw":{"emoji":"⚠️","os":["linux","darwin","win32"]}}\r
---\r
\r
# ⚠️ 已废弃:\x3CSkill Name>\r
\r
## 废弃notify\r
\r
此 Skill 已于 **YYYY-MM-DD** 正式废弃。\r
\r
### 为什么废弃?\r
\x3C原因>\r
\r
### 替代plan\r
请使用 **\x3Cnew-skill-name>**:\r
- 链接:clawhub install \x3Cnew-skill>\r
\r
### 时间线\r
- 废弃notify:YYYY-MM-DD(vX.Y.Z)\r
- 最后支持:YYYY-MM-DD(vX.Y.Z+1)\r
- 完全移除:待定\r
\r
### 迁移指南\r
详见 [references/migration.md](references/migration.md)\r
```\r
\r
---\r
\r
## 维护record\r
\r
### record模板(save至 `references/maintenance-log.md`)\r
\r
```markdown\r
# Skill 维护record\r
\r
## Skill 信息\r
- 名称:\x3Cname>\r
- 当前版本:\x3Cversion>\r
- 维护者:\x3Cmaintainer>\r
\r
## 维护历史\r
\r
### 维护 #N — YYYY-MM-DD\r
\r
**类型**:[Bug修复/Function增强/security补丁/废弃/其他]\r
**版本**:\x3Cold> → \x3Cnew>\r
**变更摘要**:\x3Csummary>\r
\r
#### 变更详情\r
\x3Cdetailed changes>\r
\r
#### securityreview\r
- CVSS:\x3Cscore>\r
- 结论:[通过/reject/条件通过]\r
\r
#### 测试结果\r
- [ ] 测试通过\r
\r
#### publish信息\r
- publish日期:YYYY-MM-DD\r
- ClawHub 版本:\x3Cversion>\r
```\r
\r
---\r
\r
## 快速参考\r
\r
### trigger命令\r
\r
| 用户请求 | execute动作 |\r
|---------|---------|\r
| "修复 Skill XX 的 bug" | 诊断 → analyze → implement → securityreview → publish |\r
| "为 Skill XX 增加 XX Function" | 需求confirm → analyze → implement → securityreview → publish |\r
| "upgrade Skill XX 的依赖" | 依赖检查 → 兼容性analyze → update → securityreview → publish |\r
| "discover Skill XX 有security漏洞" | 🚨 紧急通道 → 立即隔离 → 紧急修复 → 紧急publish |\r
| "废弃 Skill XX" | 废弃assess → notify用户 → publish废弃版本 → 保留迁移指南 |\r
\r
### 常见错误\r
\r
1. **版本号错误**:Bug 修复用 MAJOR upgrade → 应为 PATCH\r
2. **跳过securityreview**:紧急修复未做securityreview → 必须补审\r
3. **不update changelog**:变更未record → 版本历史不完整\r
4. **Breaking Change 未notify**:未inform用户 → 用户upgrade后Function损坏\r
5. **废弃 Skill 未提供替代**:用户无法迁移 → 影响用户体验\r
\r
---\r
\r
## 版本历史(Changelog)\r
\r
| 版本 | 日期 | Changes | 审核人 |\r
|------|------|---------|--------|\r
| **1.1.0** | 2026-04-13 | 新增 Agent 调用接口层(Inter-Agent Interface):7个 Task 类型(diagnose/patch/security-patch/deprecate/emergency-isolate/health-check/dependency-audit);CVE 紧急通道 SLA system;emergency-isolate authorizeverify;与 ai-skill-creator / ai-skill-optimizer 接口关系Definition | CTO-001 / CISO-001 |\r
| **1.0.0** | 2026-04-11 | Initial version:CTO 版本govern5步维护process + CISO security运营standard(漏洞respond SLA + 补丁manage)+ 废弃manageprocess | CTO-001 / CISO-001 |\r
\r
## rollbackstrategy(Rollback)\r
\r
> 如维护操作失败,execute以下steprecover:\r
\r
```bash\r
# recover到上1个可用版本\r
git checkout tags/v\x3C上1版本> -- SKILL.md scripts/ references/\r
\r
# verifyrollback成功\r
git log --oneline -3\r
```\r
\r
**rollbacktrigger条件**:\r
- `emergency-isolate` 后:满足 CVE 已修复 + CISO-001 复审通过 + CQO-001 验收通过后方可解除隔离\r
- `patch` 失败:rollback到隔离前版本,notify CTO-001\r
- `deprecate` 误操作:recover `deprecated: false`,notify CRO-001\r
\r
**解除 emergency-isolate 条件**:\r
1. CVE 已修复(CVSS \x3C 7.0)\r
2. CISO-001 security复审通过\r
3. CQO-001 质量验收通过\r
4. CTO-001 书面authorize解除隔离\r
Usage Guidance
This document is coherent for a skill-maintainer workflow, but it relies on the host platform to enforce isolation and caller identity. Before installing or invoking it, confirm that: 1) the platform enforces caller authentication/authorization (CISO-001/CTO-001 cannot be spoofed), 2) actions that modify skill files require audited, authenticated sessions, 3) emergency-isolate and security-patch are gated by multi-party authorization or at least immutable audit logs, and 4) session isolation (sessions_send/sessions_spawn) is actually supported. If those platform controls are not present, treat the workflow as risky because it permits automated file modifications and emergency isolation based on simple string fields.
Capability Analysis
Type: OpenClaw Skill Name: ai-skill-maintainer Version: 1.1.0-en2 The ai-skill-maintainer skill is a high-privilege administrative tool designed to manage, patch, and publish other AI skills. It possesses risky capabilities including shell execution for packaging/publishing (clawhub) and version control (git) as seen in SKILL.md. While the instructions include security-conscious features like path injection prevention and CVE response SLAs, the skill's reliance on prompt-based authorization checks (e.g., verifying 'CISO-001' or 'CTO-001' identities) for sensitive tasks like 'emergency-isolate' and 'security-patch' represents a significant attack surface for prompt injection. No evidence of intentional malice or data exfiltration was found, but the broad permissions required for its stated purpose warrant a suspicious classification.
Capability Assessment
Purpose & Capability
The name/description describe a skill-maintainer workflow and the SKILL.md contains interfaces and task types (diagnose, patch, security-patch, deprecate, isolate, dependency-audit) that align with that purpose. Declared permissions (read/write skills directory, L3) and the tasks that modify skill files are appropriate for a maintainer role.
Instruction Scope
Instructions specify task schemas, an isolated calling model, and simple pseudo-code for authorization. They reference modifying skill files and reading dependency lists (dependencies[], cves[]) which are expected for this workflow. However, the document assumes platform-level enforcement of caller identity/authorization and isolation (e.g., sessions_send/sessions_spawn, L3 permissions) but does not provide strong cryptographic or unforgeable authorization mechanisms — authorization fields are simple strings (CISO-001/CTO-001) and could be spoofed if the platform does not enforce them. This is a design caveat, not an incoherence with stated purpose.
Install Mechanism
No install spec or code files — instruction-only skill. This is lowest-risk from an install perspective and consistent with being a governance/workflow document.
Credentials
No required environment variables, binaries, or config paths are listed. The only implied resource is the ability to read/write the skills directory (L3 permission), which is proportionate to a maintainer workflow.
Persistence & Privilege
always is false and the skill is user-invocable (defaults). The skill's tasks include operations that can change skill files and isolate skills; this is expected for a maintainer. Because the SKILL.md expects emergency isolation and high-priority actions, verify the platform enforces strict audit logging and that only authorized principals can trigger emergency-isolate or security-patch tasks.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install ai-skill-maintainer
  3. After installation, invoke the skill by name or use /ai-skill-maintainer
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.1.0-en2
Full body English translation
v1.1.0-en
English version
v1.1.0
v1.1.0: 新增Agent调用接口层,7个TASK类型,CVE紧急通道SLA,emergency-isolate授权验证,health-check健康阈值,changelog和回滚策略
Metadata
Slug ai-skill-maintainer
Version 1.1.0-en2
License MIT-0
All-time Installs 1
Active Installs 0
Total Versions 3
Frequently Asked Questions

What is AI Skill Maintainer (EN)?

AI公司 Skill 维护工作流(CTO 版本govern + CISO security运营standard版)。当需要对已publish的 Skill 进行版本update、bug修复、Function增强、依赖upgrade、security补丁、废弃(deprecation)manage时trigger。... It is an AI Agent Skill for Claude Code / OpenClaw, with 126 downloads so far.

How do I install AI Skill Maintainer (EN)?

Run "/install ai-skill-maintainer" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is AI Skill Maintainer (EN) free?

Yes, AI Skill Maintainer (EN) is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does AI Skill Maintainer (EN) support?

AI Skill Maintainer (EN) is cross-platform and runs anywhere OpenClaw / Claude Code is available (linux, darwin, win32).

Who created AI Skill Maintainer (EN)?

It is built and maintained by JohnSmithfan (@johnsmithfan); the current version is v1.1.0-en2.

💬 Comments