repo-tour
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted data from external GitHub repositories, creating a surface for indirect prompt injection. 1. Ingestion points: README.md, build manifests (e.g., package.json, Cargo.toml), source code entry points, pull requests, and issue bodies (SKILL.md). 2. Boundary markers: Absent. The instructions do not define delimiters or specific warnings to ignore instructions embedded within the fetched repository content. 3. Capability inventory: Read-only access via the GitHub MCP tool; the skill does not use tools for file modification or code execution. 4. Sanitization: Absent. Content fetched from the remote repository is used directly to synthesize the markdown report.
Audit Metadata