sf-intake
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Node.js scripts located in the .specforge/core/scripts/ directory to manage the work item lifecycle. Specifically, it uses status.mjs to check work item counts, create-work.mjs to initialize new tasks, and instructions.mjs to verify workflow configurations.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes untrusted user data such as new requests and bug reports. 1. Ingestion points: User-provided descriptions for new work items and project wiki files. 2. Boundary markers: No explicit delimiters defined in instructions. 3. Capability inventory: Execution of local scripts via node as defined in commands.json. 4. Sanitization: No instructions provided for sanitizing input before use in CLI arguments.
Audit Metadata