go-dev
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill directs the agent to utilize repository-specific runners and existing scripts for core development tasks such as building, linting, and testing. This creates a vector for executing potentially malicious code if the repository scripts are untrusted or have been tampered with.
- Evidence:
references/core-go-rules.md: "Use the repository runner (just, make, or existing scripts) for build, lint, and test commands." - [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to analyze and adopt patterns from the repository and sibling services. This ingestion of external data lacks explicit boundary markers or sanitization, which could allow malicious instructions hidden within the codebase (e.g., in comments or metadata) to influence the agent's actions.
- Ingestion points: Local repository files and sibling service codebases.
- Capability inventory: File system modification, command execution via repository runners.
- Boundary markers: None identified in the instructions.
- Sanitization: None identified in the instructions.
Audit Metadata