research
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from untrusted external sources, including official documentation, source repositories, and user-supplied artifacts. This creates a surface for indirect prompt injection where malicious instructions embedded in researched materials could influence the agent's behavior.
- Ingestion points: Repository source files, logs, migration scripts, tickets, and external documentation (SKILL.md).
- Boundary markers: None specified in the provided instructions.
- Capability inventory: Reading local files, executing local commands or tests (SKILL.md).
- Sanitization: None specified in the provided instructions.
- [COMMAND_EXECUTION]: The skill explicitly instructs the agent to 'Run small local commands or tests when they can verify repo behavior cheaply'. This provides a mechanism for the agent to execute shell commands on the host system as part of its research process.
Audit Metadata