workflow-end
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands and scripts to manage workflow lifecycles. * Evidence: Invokes
node .claude/hooks/lib/review-receipt.cjs,node .claude/scripts/lib/workflow-baseline.cjs, andpython3 .claude/scripts/code_graph. * Evidence: Usesgit diffandgit ls-filesto detect business logic changes requiring test coverage. - [DYNAMIC_EXECUTION]: The skill triggers the execution of Node.js and Python scripts located within the project's
.claude/directory at runtime. * Evidence: References automated scripts in.claude/scripts/lib/for generating baseline reports and performing cleanup. - [INDIRECT_PROMPT_INJECTION]: The skill processes external workflow manifests and review reports, which serves as a potential vector for indirect prompt injection. * Ingestion points: Reads Tier-2 manifests,
review-reportfiles, andplan.mdchecklist items inSKILL.md. * Boundary markers: Absent for raw file ingestion, although the instructions mandate citing specificfile:lineevidence for all claims. * Capability inventory: Includes shell command execution (node,python3,git) and file-system access within project and temporary directories. * Sanitization: No explicit sanitization or structural validation of the ingested report content is defined.
Audit Metadata