coach-interview

Pass

Audited by Gen Agent Trust Hub on Jul 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill explicitly instructs the agent to execute a local script, scripts/ask_cited.py, to fetch information. This includes a shell loop to pre-fetch multiple topics at once.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing untrusted data.
  • Ingestion points: The skill requires the user to provide data/strong_workouts.csv, which is then used for analysis and to build the cited protocol.
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content embedded within the CSV file.
  • Capability inventory: The agent has the capability to execute shell commands (python3 scripts/ask_cited.py) and write to the local filesystem (private/protocol.md).
  • Sanitization: The instructions do not mention any validation or sanitization of the CSV data before it is processed or used to generate queries for the citation script.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 19, 2026, 12:09 PM
Security Audit — agent-trust-hub — coach-interview