familiarize
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed for read-only repository exploration and uses local bash commands and Serena MCP tools for orientation. It does not exhibit malicious patterns such as exfiltration or privilege escalation.
- [DATA_EXPOSURE]: The skill reads project configuration files like .envrc and .env.example (Step 4). While these files can contain sensitive configuration, the skill has no network capabilities to exfiltrate this data and only uses it to provide the user with a project overview.
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from documentation files. 1. Ingestion points: Reads CLAUDE.md, README.md, and files in .llmdocs/ (Steps 3, 4). 2. Boundary markers: None identified. 3. Capability inventory: Local bash execution (find, ls, git) and Serena LSP tools (get_symbols_overview, find_symbol). 4. Sanitization: None observed. The risk is limited by the lack of write or network permissions.
- [COMMAND_EXECUTION]: The skill executes hardcoded bash commands (find, ls, git) to map directory structures and retrieve repository metadata. These commands are restricted to local discovery and do not accept untrusted input that could lead to injection.
Audit Metadata