team-research

Pass

Audited by Gen Agent Trust Hub on Nov 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from codebase files and agent outputs, which represents a surface for indirect prompt injection. Malicious instructions embedded in researched codebase files could attempt to hijack agent behavior. This risk is addressed by an explicit instruction to disregard any instructions found within external returns and to strictly adhere to host-defined external data rules.
  • [COMMAND_EXECUTION]: The skill executes local shell scripts (discover-topic.sh) and Node.js scripts (render-research.mjs) located within its own directory structure. These scripts use user-provided directory paths as arguments. The implementation utilizes double-quoting in shell commands to prevent common command injection vectors, and the execution is restricted to the skill's internal scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 15, 2026, 12:30 PM