Back to skill

Security audit

doc-writer

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this documentation-writing skill suspicious because it reports unsafe setup and handoff guidance; the evidence is incomplete because Skillspector scanned 22 of 25 components and had partial analysis.

Review the setup and handoff instructions before installing. Avoid following the environment copy step if `.env.local` already contains settings, and check that handoffs will redact secrets and unnecessary personal information. Treat the verdict cautiously: the overall confidence is medium, and analysis was partial, with 22 of 25 components scanned and some static checks limited.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · templates/architecture/DEV-STANDARDS.md (reported line 47)May include surrounding context.

md
- 提交规范:〈如 Conventional Commits:`feat|fix|docs|refactor|test|chore(scope): subject`〉
- 合并纪律:〈如 squash 合并、主干禁 force-push、变更至少 1 人审查〉
- CI 门禁:〈lint + test + build 全绿才可合并〉——门禁以流水线配置文件(〈.ci/pipeline.yaml〉)为准,本节与配置不一致时修本节
- 环境约定:〈本地/预发/生产配置来源(.env / 配置中心);密钥不入库,.env.example 仅含键名与样例值〉

## 3. 编码规范索引(按栈)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · templates/architecture/DEV-STANDARDS.md (reported line 67)May include surrounding context.

bash
# 环境:〈Node.js 20 LTS / pnpm 9〉
pnpm install                  # 安装全部依赖(workspace 递归)
cp .env.example .env.local    # 配置本地环境变量(必填项见 .env.example)
pnpm dev                      # 本地启动(默认 <redacted-url>
pnpm test                     # 全量测试
pnpm --filter order-api test  # 单包测试

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · templates/architecture/API-CONTRACT.md (reported line 76)May include surrounding context.

  • 示例(可复制粘贴,含认证头):
bash
curl -X POST "$BASE/api/v1/orders" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $UUID" \

The environment setup command can overwrite existing local configuration without warning.

Medium
Category
Output Handling
Confidence
97% confidence
Finding

The command cp .env.example .env.local overwrites an existing local configuration without checking or requesting confirmation. Because this template explicitly anticipates literal execution by coding agents, following the setup instructions can destroy customized settings or locally stored credentials, although no disclosure is demonstrated.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · templates/delivery/DEPLOY-GUIDE.md (reported line 45)May include surrounding context.

6. 启动

〈命令〉

7. 验证

curl # 预期输出:〈状态码/响应体〉

text

## 3. Docker 方式(如提供)

Requires verbatim retention and propagation of user input without a privacy warning or sensitive-data handling exception.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The template unconditionally requires verbatim preservation and downstream propagation of user input in archived handoff documents, with no exception for credentials, personal information, or confidential content. In a documentation workflow, this can persist sensitive material in repository history and expose it to additional roles or agents beyond its intended audience.

Content

No source excerpt is available for this finding.