grooming

Pass

Audited by Gen Agent Trust Hub on Sep 30, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided stories and acceptance criteria, which are untrusted data sources. This creates a surface for indirect prompt injection where malicious instructions could be embedded in the requirements text to influence the agent's actions.
  • Ingestion points: User-provided stories and acceptance criteria processed in SKILL.md.
  • Boundary markers: The skill does not define explicit delimiters to isolate untrusted data from processing instructions.
  • Capability inventory: The skill writes to markdown files in 'docs/business/' and executes a Python script via the shell.
  • Sanitization: No sanitization or input validation is performed on the ingested text.
  • [COMMAND_EXECUTION]: The skill invokes a local Python script using 'python3 /scripts/build_ac_index.py' to regenerate indices after updating requirements files.
  • Evidence: Execution command found in the block of SKILL.md.
  • Note: The use of a directory placeholder implies reliance on a pre-configured local tool environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 30, 2026, 02:49 AM