The Agent Skills Directory

[SAFE]: The skill's primary script, detect_plugins.mjs, performs local file system operations to identify WordPress plugins. It uses safe path handling and limits the depth and number of files processed to avoid performance issues.
[COMMAND_EXECUTION]: The skill executes local scripts via Node.js to triage the project and detect plugins. These operations are limited to the repository context and do not involve shell injection or untrusted input execution.
[EXTERNAL_DOWNLOADS]: All external links point to the official WordPress Developer Resources (developer.wordpress.org), which is a trusted and authoritative source for documentation.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingests metadata from PHP files within the target repository. However, this is a standard risk for development tools and is mitigated by the script's focus on structured header parsing.
Ingestion points: scripts/detect_plugins.mjs reads the content of PHP files in the repository.
Boundary markers: The script outputs structured JSON; however, the agent's final prompt composition logic is not defined here.
Capability inventory: The agent is granted bash and node capabilities for development tasks.
Sanitization: Content is extracted via regex and trimmed, but no LLM-specific sanitization is applied to the extracted metadata.

wp-plugin-development