orchestrator
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to manage a git-based workflow, including
git worktree,git rebase,git cherry-pick, and project-specific build systems for 'byte gating' and 'rule checking'. - [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface by ingesting and merging code and metadata produced by 'matcher workers' (sub-agents).
- Ingestion points: Worker output, lane branches, and the
AGENTS.mdconfiguration file. - Boundary markers: The instructions state that
AGENTS.mdtakes precedence and requires reporting per-function status, but lacks explicit sanitization instructions for worker-provided code. - Capability inventory: The skill performs file-write operations (merging), git operations, and executes build scripts on the integrated code.
- Sanitization: There is no explicit requirement to sanitize or escape the content generated by workers before it is integrated into the build process.
Audit Metadata