verify
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs extensive shell command execution for toolchain discovery, git operations, and running test/linter suites (Phases 1, 3, 4, and 7). This capability depends on the integrity of the project's local environment and configuration files like 'package.json', 'Makefile', or 'Cargo.toml'.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository data, including source code, diffs, and external files like '--plan-file' or '--epic-context', and interpolates this content into prompts for various sub-agents.\n
- Ingestion points: Scoped files, git diffs, '--plan-file', and '--epic-context' files (Phases 1, 5, and 6).\n
- Boundary markers: Utilizes logical delimiters such as 'VERIFICATION SCOPE' and 'CONTEXT_BUNDLE', but lacks explicit sanitization or strong instructions for agents to disregard embedded commands within the analyzed data.\n
- Capability inventory: Includes shell command execution (Phases 3/4/7), sub-agent invocation (Phase 6), and automated file modification in 'auto-fix' mode (Phase 9).\n
- Sanitization: The skill does not specify any sanitization or escaping of the ingested data before it is interpolated into prompts for the sub-agents.
Audit Metadata