The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill performs SEO audits on external content, creating a vector for indirect prompt injection if a processed URL or file contains malicious instructions aimed at the agent.
Ingestion points: Data is ingested through the /seo hreflang audit <directory-or-url> command defined in SKILL.md.
Boundary markers: There are no specific instructions or delimiters mentioned to isolate the external content from the agent's command instructions.
Capability inventory: The skill involves network reads (fetching URLs), file system reads (scanning directories), and file generation (creating hreflang-sitemap.xml).
Sanitization: No explicit sanitization or filtering of external input is described in the skill's logic.
[NO_CODE]: The skill consists entirely of markdown instructions and reference text files; it does not include any executable scripts, binaries, or complex configuration files that could hide malicious logic.

seo-hreflang