Skills
Navigate
skills/victorzhuk/skills/z-to-prd/Gen Agent Trust Hub

z-to-prd

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from existing project documents and conversation logs to generate a PRD.\n
  • Ingestion points: Data is pulled from docs/design/<slug>-brief.md, the conversation history (referencing z-grill-with-docs), and codebase modules.\n
  • Boundary markers: Absent. No delimiters or instructions are provided to the agent to distinguish between the skill's instructions and potentially malicious content within the source documents.\n
  • Capability inventory: The skill writes to the file system at docs/prd/<slug>.md and suggests executing shell commands for verification.\n
  • Sanitization: Absent. There is no mention of validating or escaping content from the ingested files before using it to generate the PRD.\n- [COMMAND_EXECUTION]: The skill instructs the agent to use the fd command in the verification step to confirm the existence of the generated PRD file in the docs/prd/ directory.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 09:30 AM