The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches the @rsdoctor/agent-cli package and version metadata from the official npm registry (registry.npmjs.org). It also contains instructions for downloading package tarballs to inspect type definitions. These operations target a well-known service and are standard for development tooling.
[COMMAND_EXECUTION]: The skill executes shell commands for package management (npm install), environment verification (rsdoctor-agent --version), filesystem searching (rg), and triggering project build processes (pnpm run build). These actions are core to the skill's purpose and include instructions to only run in authorized environments.
[INDIRECT_PROMPT_INJECTION]: The skill processes rsdoctor-data.json, an external build artifact, to generate recommendations. This creates a vulnerability surface where malicious data could influence agent behavior.
Ingestion points: Reads rsdoctor-data.json from various relative project paths (dist/, output/, etc.) as defined in SKILL.md.
Boundary markers: No explicit delimitation or instruction-filtering is applied to the ingested JSON content.
Capability inventory: Includes execution of CLI analysis tools and build scripts as documented in references/command-map.md.
Sanitization: No specific sanitization or validation of the JSON fields is described prior to processing.

rsdoctor-analysis