workflow-feature
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill serves as an orchestration layer for multi-step development tasks, creating an attack surface where untrusted inputs could potentially influence agent actions.
- Ingestion points: The workflow processes user prompts through the
$start-workflowtool and reads external configuration files such as.claude/workflows.jsonand various project specification documents. - Boundary markers: The skill does not define explicit delimiters or 'ignore' instructions to mitigate the risk of the agent obeying instructions embedded within the processed project files or user-provided design links.
- Capability inventory: The skill is authorized to spawn sub-agents (
spawn_agent), execute local Node.js scripts, and perform extensive file system operations across the project repository (e.g., indocs/,plans/, andtmp/directories). - Sanitization: No sanitization, validation, or escaping of ingested data is specified before it is used to drive subsequent workflow steps.
- [COMMAND_EXECUTION]: The skill contains an explicit instruction to execute a local Node.js maintenance script (
node .claude/scripts/lib/workflow-skill-contract.cjs --write) to synchronize documentation blocks.
Audit Metadata