The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest content from arbitrary external websites, which acts as a surface for malicious instructions to influence the agent's behavior.\n
Ingestion points: snapshot.mjs fetches HTML, CSS, and JS from user-provided URLs entering the agent context.\n
Boundary markers: No explicit boundary markers are used in the automated processing of external content.\n
Capability inventory: The skill has the ability to execute shell scripts, write files to the repository, and make network requests.\n
Sanitization: tojsx.mjs removes <script> tags and event handlers (e.g., onclick) during the conversion to JSX, reducing the risk of executing malicious code embedded in the source HTML.\n- [EXTERNAL_DOWNLOADS]: Downloads assets and standard developer tools from known sources.\n
Mirrored assets (images, fonts, media) are downloaded from the source website and well-known CDNs like Google Fonts, Cloudflare, and Supabase.\n
Standard NPM packages required for functionality are installed during setup.\n- [COMMAND_EXECUTION]: The agent runs multiple local utility scripts using node to orchestrate the migration, including detection, snapshots, and code conversion.\n- [REMOTE_CODE_EXECUTION]: Executes a standard setup command npx playwright-core install chromium which downloads a browser binary from an official source.

migrate