blueprint-prp-create

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash tool for local project state management and utility operations.
  • It executes 'test' and 'find' to verify the existence of the blueprint directory structure and locate existing documentation.
  • It uses 'jq' and 'awk' to read and increment sequential ID counters stored in 'docs/blueprint/manifest.json'.
  • [EXTERNAL_DOWNLOADS]: The skill uses 'WebSearch' and 'WebFetch' to gather research data from the internet.
  • It retrieves official documentation, community discussions (Stack Overflow, GitHub), and best practices to inform the product requirement prompt.
  • [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it processes untrusted data from the web and incorporates it into a prompt packet (PRP) intended for later execution.
  • Ingestion points: External data retrieved via 'WebSearch' and 'WebFetch' in Step 3.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the fetched content are defined in the workflow.
  • Capability inventory: The skill utilizes 'Write' for file creation and 'Bash' for local utility execution.
  • Sanitization: There is no mention of sanitizing or escaping the content fetched from the web before it is written into the PRP Markdown document.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 01:04 AM