The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute several local Node.js utility scripts (e.g., doctor.mjs, instructions.mjs, quality-suite.mjs, and artifact-graph-status.mjs) found within the project's .specforge/core/scripts/ directory to manage and verify the development workflow state.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface because it reads and processes data from various markdown files within the user's repository to determine its logic flow.
Ingestion points: The agent is instructed to read contents from .specforge/registry.yaml, 00-intake/brief.md, 00-intake/prd.md, and other specification files to identify status markers like [NEEDS DECISION] or specific gate results.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat the content of these external files as untrusted or to ignore any embedded directives within them.
Capability inventory: The skill has the capability to execute shell commands via the Node.js scripts referenced in the workflow.
Sanitization: No validation or sanitization of the markdown file content is performed; the agent is simply directed to search for specific strings or patterns to make routing decisions.

sf-router