The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill performs SEO audits by scanning and ingesting content from external URLs or local directories provided by the user via the /seo hreflang audit command. This creates a surface for indirect prompt injection, where malicious instructions hidden in the audited site's metadata or HTML could potentially influence the agent's behavior.
Ingestion points: Target URLs and local content directories provided as arguments to the audit command.
Boundary markers: The skill instructions do not specify the use of clear delimiters or instructions to ignore embedded commands within the audited data.
Capability inventory: The skill utilizes network fetching to scan external sites and has the ability to execute a local Python script for processing.
Sanitization: There is no evidence of content sanitization or filtering before the data is processed by the agent.
[DYNAMIC_EXECUTION]: The skill instructions include a reference to an external execution pattern where content quality scoring is delegated to a Python script located at a specific path.
Evidence: The file references/machine-translation-qa.md describes running content_quality.py via the path "${CLAUDE_PLUGIN_ROOT}/scripts/claude-seo".
Context: This script is a vendor-provided resource managed within the skill's plugin environment, used for entropy and AI-pattern detection.

seo-hreflang