milestone
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill instructions are focused on standard software development lifecycle tasks and do not exhibit malicious intent, obfuscation, or unauthorized access to sensitive system resources.
- [COMMAND_EXECUTION]: The workflow describes standard development commands such as branching, implementing code changes, and running tests. These operations are logically restricted to the user-specified repository and milestone context.
- [PROMPT_INJECTION]: No attempts to override agent behavior, bypass safety filters, or extract system prompts were detected in the skill instructions.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from GitHub issues and pull requests, creating a potential attack surface. However, the risk is mitigated by explicit requirements for human review and multi-agent verification. (1) Ingestion points: GitHub issue descriptions, comments, and project state metadata. (2) Boundary markers: None. (3) Capability inventory: File system modification, subprocess execution for testing, and GitHub API interactions. (4) Sanitization: Not explicitly mentioned for external text content.
Audit Metadata