insecure-source-code-management

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill explicitly documents techniques and tools to locate and extract VCS/backup artifacts and secrets (e.g., .git, .env) which enable data exfiltration and credential theft if used outside authorized assessments.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). SKILL.md instructs runtime HTTP probing of exposed repo/backup paths (e.g., /.git/HEAD, /.git/config, /.svn/entries, /.env, /.DS_Store), whose recovered contents are outside-authored server free text and would be read into the agent’s LLM context if the workflow ingests fetched responses.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly tells operators to "run" external recovery tools (e.g., arthaud/git-dumper, internetwache/GitTools, WangYihang/GitHacker) which are external repositories/tools executed at runtime and therefore represent runtime dependencies that execute remote code.