The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands to build and deploy the blog. It relies on user-provided variables BUILD_CMD and DEPLOY_CMD to run these processes (e.g., python3 build.py, vercel deploy --prod).
[DYNAMIC_EXECUTION]: The skill dynamically determines which commands to execute based on external configuration files (kit.toml, .env) or arguments, which could lead to arbitrary command execution if the configuration is malicious.
[INDIRECT_PROMPT_INJECTION]: The skill processes markdown content from local files and verifies deployment by fetching and parsing HTML from a live website, creating an attack surface for instructions embedded in external content.
Ingestion points: Reads markdown post files from POSTS_DIR and fetches live page content from $BLOG_URL using curl.
Boundary markers: None identified in the instructions to delimit or ignore instructions within the processed data.
Capability inventory: Includes shell execution (build/deploy), file reading, and network requests (curl).
Sanitization: No explicit sanitization or validation of the content fetched from the live URL is described before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill performs network operations using curl to verify the deployment status at the configured BLOG_URL. It also references deployment tools from well-known services including Vercel, Netlify, Cloudflare, and Railway.

blog-publish