37-micro-interactions
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill contains instructions that direct the agent to ingest and process data from the user's local repository, which serves as a potential vector for malicious instructions embedded in codebase files.
- Ingestion points: The implementation contract in SKILL.md requires the agent to 'Inspect the existing repository, framework, dependencies, and conventions' before implementation.
- Boundary markers: There are no explicit markers or instructions provided to the agent to treat repository content as untrusted or to ignore instructions found within analyzed files.
- Capability inventory: The skill documentation itself does not define any scripts, network operations, or privileged file system writes; it relies on the host agent's native capabilities.
- Sanitization: The skill does not specify any sanitization, filtering, or validation steps for the content discovered within the repository.
Audit Metadata