execute
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONMETADATA_POISONING
Full Analysis
- [COMMAND_EXECUTION]: The skill executes several bash scripts and git commands to manage worktrees, branches, and verify task states. These are intended for project management and implementation workflows.
- [METADATA_POISONING]: The
SKILL.mdfrontmatter includes references to various books on software engineering and delivery. These are used as sources for the workflow's principles and do not contain malicious instructions. - [COMMAND_EXECUTION]: The skill uses
gh api graphqlto query GitHub issue status, blockers, and PR references to ensure a task is ready for implementation. The logic includes a verdict system to prevent duplicate work or starting blocked tasks. - [COMMAND_EXECUTION]: The
references/sidecar.shscript manages a private sidecar repository for project state, research archives, and configuration. It performs git operations within that isolated directory to persist pack-authored content without polluting the client repository.
Audit Metadata