The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and analyze untrusted data from external web pages (HTML, robots.txt, sitemaps, and llms.txt).
Ingestion points: The workflow in SKILL.md directs the agent to fetch content from user-provided URLs using curl and retrieve metadata files from remote domains.
Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to disregard instructions potentially embedded within the target website's HTML or llms.txt file.
Capability inventory: The skill uses curl for network operations and runs a local Node.js script (resources/score.js) to process the audit results.
Sanitization: No sanitization or filtering of the fetched external content is described before the agent processes it against the scoring rubric.
[COMMAND_EXECUTION]: The skill utilizes shell commands to perform its core tasks.
Evidence: SKILL.md instructs the agent to execute curl -A "GPTBot" <url> for fetching content and node resources/score.js for calculating scores. While standard for this use case, these commands represent the skill's operational surface within the host environment.

seo-geo-aeo-audit