The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python helper script and the Vercel CLI to process deployment data. This behavior is restricted to the skill's internal files and established tools.
[EXTERNAL_DOWNLOADS]: Network requests are made via curl to verify the availability of Vercel preview URLs. These operations are directed at a well-known service provider and align with the stated purpose of monitoring deployments.
[PROMPT_INJECTION]: The skill processes external data sources like terminal logs and command outputs. While this creates an indirect injection surface, the processing is limited to extracting specific patterns via regular expressions, minimizing the risk of instruction override. Ingestion points: terminal output and pasted logs via scripts/extract-vercel-links.py. Boundary markers: not explicitly defined. Capabilities: vercel cli execution, curl probing, and automation creation. Sanitization: regex-based extraction of URLs and statuses.

vercel