The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project source code, configurations, and manifests, which serves as an ingestion point for potentially malicious instructions embedded in data.
Ingestion points: The agent is directed to read from packages/*/*/src, apps/, python/, and runtime scripts.
Boundary markers: No specific delimiters or instructions to ignore embedded natural language commands are provided for the analyzed content.
Capability inventory: The agent uses search tools like rg and is empowered to write design records (Agent Notes) and update documentation files such as READMEs and JSDoc.
Sanitization: There are no explicit sanitization steps for the ingested repository content.
Mitigation: The instructions explicitly separate the 'survey' phase from implementation, requiring user authorization and evidence before any changes are applied.
[COMMAND_EXECUTION]: The skill instructs the agent to run local development and validation commands.
Evidence: Commands include pnpm run doc-sync, pnpm run lint, and git diff --check.
Context: These are standard utility commands for a developer environment and are used to validate documentation and code quality after the analysis phase.

dsh-find-simplifications