The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute curl commands for verifying the availability of external image URLs fetched from the web. Specifically, it instructs the agent to run curl -sI "url" | head -1 to check for HTTP 200 responses. This pattern incorporates externally sourced strings into a shell pipeline.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external websites to gather statistics and citations, creating a vulnerability surface. * Ingestion points: WebSearch and WebFetch results are ingested in SKILL.md (Phase 2) to find statistics and images. * Boundary markers: The instructions do not define clear delimiters or markers to separate untrusted external content from the agent's core instructions. * Capability inventory: The skill utilizes Bash, Write, Edit, and WebFetch tools in SKILL.md, providing a significant capability surface if the agent is manipulated by ingested content. * Sanitization: There are no explicit instructions to sanitize or validate the content of the fetched statistics or website data before incorporating them into the article or using them in shell commands.
[EXTERNAL_DOWNLOADS]: The skill fetches images and data from established external services including Pixabay, Unsplash, and Pexels. These operations are documented neutrally as they are aligned with the skill's primary function of blog post generation.

blog-write