The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides bash scripts intended for execution by the agent to perform directory traversal and search for architectural patterns. These scripts utilize standard system utilities including dirname, basename, grep, and jq.
[DATA_EXPOSURE]: The skill accesses local project metadata (plugin.json) and documentation files (commands/ and agents/ directories) to identify integration points and evaluate the separation of concerns. This access is restricted to the local project workspace.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from external markdown files and JSON metadata to generate architectural reports. While this establishes an ingestion surface for untrusted data, the processing is limited to structural analysis and pattern matching.
Ingestion points: Reads plugin.json and markdown files located in commands/ and agents/ directories (SKILL.md).
Boundary markers: The skill does not define specific delimiters for the ingested content but processes it through targeted grep and jq commands.
Capability inventory: Utilizes bash, grep, and jq for file processing; includes example implementation patterns using curl for local API checks.
Sanitization: Employs jq for structured data extraction, which mitigates simple injection risks during metadata parsing.

plugin-integration-checker