The Agent Skills Directory

[COMMAND_EXECUTION]: The skill employs standard shell utilities to inspect the local development environment. Evidence includes the use of lsof for port checks, grep for configuration parsing, cat for reading local status files, and curl for localhost health verification.- [EXTERNAL_DOWNLOADS]: The skill recommends ensuring local environment readiness by managing dependencies. Evidence includes instructions to execute npm install when required packages or the node_modules directory are absent.- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest and act upon data from external task files. Ingestion points: tasks/next_task.md and tasks/progress.txt. Boundary markers: Absent; data is read and interpreted directly. Capability inventory: Includes npm install, npx tsc, and load_skill calls. Sanitization: Absent; the agent reads external content without specific filtering or delimiter protection.

pre-flight-checklist