The Agent Skills Directory

[COMMAND_EXECUTION]: The skill frequently uses the Bash tool to execute diagnostic commands, including build scripts (npm run build), TypeScript compilation (npx tsc), and Cloudflare Worker management (npx wrangler). This is standard functionality for the intended use case but grants the agent significant system interaction capabilities.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) due to its core function of analyzing external, potentially untrusted data.
Ingestion points: The skill reads local source code, build logs, git history, and environment configuration files (.env.local).
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are used when the agent processes the content of these files.
Capability inventory: The agent has access to Write and Edit for file modification, Bash for command execution, and mcp__playwright for browser automation.
Sanitization: While the diagnose.sh script masks values in .env.local using sed, other data sources like source code comments or log files are processed without sanitization.

fullstack-debugger