task-execution-simple
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a standard software development process including git branch management and implementation verification. It includes safety gates such as mandatory explicit user confirmation for destructive operations including branch deletion, worktree removal, and discarding work.\n- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because its core workflow involves following instructions from local documentation files.\n
- Ingestion points: Reads implementation details, specs, and plans from
spec.md,plan.md, and files withindocs/tasks/.\n - Boundary markers: The instructions do not define specific delimiters or security guidelines to prevent the agent from obeying instructions embedded within these repository files.\n
- Capability inventory: The skill uses git for repository management (branches, worktrees, merges), modifies the local file system, and executes shell-based automated tests.\n
- Sanitization: Content from the ingested documentation is not sanitized or validated before being used to guide agent actions.
Audit Metadata