framework-docs-researcher
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, untrusted sources including web search results, official documentation sites, and GitHub issues/discussions. This creates an attack surface for indirect prompt injection where malicious content in those sources could attempt to influence the agent's behavior. However, the instructions prioritize official sources and require synthesis of information, which serves as a natural mitigation.
- Evidence:
- Ingestion points: Web searches, GitHub issues/discussions, and documentation fetched via 'Context7'.
- Capability inventory: The agent has access to file reading (
Read), directory traversal (Glob), content searching (Grep), and limited shell execution (bundle show). - Sanitization: No explicit instructions for sanitizing or delimiting external content are provided in the skill.
- [COMMAND_EXECUTION]: The skill uses a shell tool to execute the
bundle showcommand. This is used specifically to locate the installation paths of Ruby gems for source code analysis. This is a standard operation in development environments and is aligned with the skill's primary function.
Audit Metadata