The Agent Skills Directory

[SAFE]: The skill's primary function is to perform a standard SEO audit of internal link structures. This is a common developer task, and the instructions provided are consistent with this purpose.
[EXTERNAL_DOWNLOADS]: The skill references established and trusted sources for SEO guidelines, including Google Search Central and Screaming Frog. These references are informational and do not involve downloading or executing untrusted code at runtime.
[COMMAND_EXECUTION]: The skill suggests using the standard curl utility to inspect HTTP headers. The command logic is restricted to diagnostic header inspection (-I) and following redirects (-L), which is safe for the intended audit workflow and does not involve arbitrary command execution.
[INDIRECT_PROMPT_INJECTION]: The skill involves processing external data by crawling website links, which is a potential injection surface. However, this is the primary purpose of the tool.
Ingestion points: Internal links extracted from the website being audited (referenced in SKILL.md and references/rule.md).
Boundary markers: None present in the instructional text.
Capability inventory: HTTP status code checking and link traversal using curl.
Sanitization: The agent is instructed to report technical metadata (status codes and URLs) rather than executing content found on the crawled pages, mitigating the risk of executing injected instructions.

redirect-chains