The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to automatically run npm run build to synchronize project assets after modifying the registration JSON file. This is a standard operational procedure for maintaining the project's build state and documentation.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes metadata from external repositories and local files. • Ingestion points: Data is retrieved from local SKILL.md files and external GitHub repositories provided by the user. • Boundary markers: No specific delimiters or instructions are implemented to prevent the agent from processing instructions that might be embedded in the ingested metadata. • Capability inventory: The skill performs file system modifications to src/data/skills.json and executes the npm run build command. • Sanitization: Ingested metadata is appended to the project's registry without explicit sanitization or validation of its content.

skill-manager