The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell scripts (scripts/start-server.sh, scripts/stop-server.sh) to orchestrate a local Node.js server process. These scripts use system utilities like ps and kill for legitimate lifecycle management and process termination.\n- [PROMPT_INJECTION]: The skill facilitates brainstorming by analyzing local project files, which may contain untrusted data.\n
Ingestion points: SKILL.md instructs the agent to check project files, documentation, and recent commits to understand context.\n
Boundary markers: No explicit delimiters or boundary markers are utilized when incorporating project data into the agent's context.\n
Capability inventory: The skill can write to the filesystem, execute local shell commands, and manage a local web server.\n
Sanitization: No explicit sanitization or filtering of ingested project content is defined before the data is processed or rendered in the visual companion browser.\n- [SAFE]: The Node.js server (scripts/server.cjs) follows security best practices for local tools by binding to 127.0.0.1 by default and using path.basename to prevent path traversal vulnerabilities when serving session-specific HTML files.

brainstorming