create-qa-list

Pass

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local Python script scripts/build_export.py to transform a structured JSON file into CSV and HTML deliverables. This is an appropriate and documented use of local execution for the skill's primary purpose.\n- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it is designed to ingest and process untrusted external data.\n
  • Ingestion points: The skill processes user-provided technical specifications, pull request diffs, and content from external URLs as specified in the workflow instructions in SKILL.md.\n
  • Boundary markers: No specific boundary markers or delimiters are defined to wrap the untrusted input data.\n
  • Capability inventory: The agent can write local files (cases.json) and execute a local Python script to generate reports.\n
  • Sanitization: No explicit sanitization or filtering of external content is mentioned, although the skill includes 'Rule 1' and 'Rule 2' to guide the agent toward safe and verified behavior when interpreting input.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 24, 2026, 01:41 AM
Security Audit — agent-trust-hub — create-qa-list