plan-review
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local project script located at
.claude/scripts/code_graphto perform dependency analysis and verify plan coverage. It uses subcommands such asquery importers_of,connections, andtraceto ensure all affected files are accounted for in the implementation plan.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes content from untrusted plan files.\n - Ingestion points: Reads
plan.mdandphase-*.mdfiles (Step 1) to extract requirements and steps.\n - Boundary markers: The instructions do not specify the use of delimiters or ignore-instructions to isolate the plan content from the review logic.\n
- Capability inventory: The skill can execute local scripts via
python, create new tasks viaTaskCreate, and write report files toplans/reports/.\n - Sanitization: No explicit sanitization or validation of the plan content is defined before processing.
Audit Metadata