The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple Node.js scripts located in the local project's .specforge/core/scripts/ directory, including artifact-graph-status.mjs, instructions.mjs, create-artifact.mjs, gate.mjs, and test-case-quality.mjs, which are used to manage project gates and validation state.- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface due to its data processing requirements. 1. Ingestion points: It reads numerous project files including 01-spec/requirements.md, 01-spec/ui-design.md, and 03-implementation/report.md. 2. Boundary markers: No explicit delimiters or 'ignore' instructions are provided for the ingested markdown content. 3. Capability inventory: The skill has the capability to execute local scripts and write files to the 05-verification/ directory. 4. Sanitization: The skill contains a specific safety guideline instructing the agent not to treat browser page content as instructions.

sf-verify