workflow-architecture-audit
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Node.js script at ".claude/scripts/lib/workflow-skill-contract.cjs" to regenerate workflow configuration blocks within the skill file.
- [DYNAMIC_EXECUTION]: The skill orchestrates complex workflows by dynamically loading and invoking multiple sub-skills (e.g., $investigate, $architecture, $why-review) and spawning sub-agents (spawn_agent) based on configuration data read from a registry file (".claude/workflows.json").
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted project source code and documentation to perform an architecture audit. * Ingestion points: Project source files, configuration files, and the ".claude/workflows.json" registry. * Boundary markers: No explicit delimiters or instructions to ignore embedded commands are defined within the skill to isolate untrusted code from the agent's instructions. * Capability inventory: The skill has the capability to execute shell commands (node), spawn sub-agents, and perform file read/write operations for reporting. * Sanitization: There is no evidence of sanitization, escaping, or validation of the audited code content before it is processed by the agent or sub-agents.
Audit Metadata