The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a shell function poll_deploy that executes curl, grep, and sleep commands to verify the state of a remote web server. This is the primary intended functionality of the skill and follows best practices for deployment verification in constrained agent environments.
[INDIRECT_PROMPT_INJECTION]: The skill implements a pattern that ingests untrusted data from a remote URL during the deployment verification process. • Ingestion points: The poll_deploy function in SKILL.md uses curl to fetch the body of a web page into a shell variable for pattern matching. • Boundary markers: There are no explicit instruction delimiters or markers to isolate the fetched content from the agent's context. • Capability inventory: The skill's execution environment includes shell access for network requests and string processing. • Sanitization: The fetched content is processed using shell case statements for string matching, which prevents the data from being directly executed as code within the polling script itself, though it remains in the agent's context.

github-pages-deploy-verification