blueprint-research
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data that could influence agent behavior if it contains malicious instructions.
- Ingestion points: The workflow ingests feature_description and tech_stack from user input, as well as files from the .agents/lessons/ directory.
- Boundary markers: Instructions for sub-tasks (e.g., toolbox-resolver, lessons-discoverer) interpolate these variables into prompts without using delimiters or instructions to disregard embedded commands.
- Capability inventory: The skill can spawn various research sub-tasks and dynamically load coding style skills.
- Sanitization: There is no evidence of sanitization or filtering of the input variables before they are used in prompts.
- [DYNAMIC_EXECUTION]: The skill dynamically loads logic during the 'Skill Injection' phase. It resolves a list of coding_styles during research and applies them to the current session using the Skill() command, allowing the agent to adapt its behavior to specific project conventions.
Audit Metadata