The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute several Python scripts to perform SEO-related tasks, including site management and findings validation.
Evidence: python "<suite-root>/scripts/validate_seo_findings.py" validate-findings <bundle>/seo-findings.json --bundle <bundle> in SKILL.md.
Evidence: python "<suite-root>/scripts/manage_markdown_site.py" build/validate/live and python "<suite-root>/scripts/manage_markdown_work.py" in references/markdown-delivery-contract.md.
[DYNAMIC_EXECUTION]: The skill uses a dynamically resolved path (<suite-root>) to locate and execute its supporting scripts. This path is determined at runtime based on environment variables (e.g., CLAUDE_PLUGIN_ROOT) or a local configuration file (.seo-suite-runtime.json).
Evidence: The resolution logic for <suite-root> is described in SKILL.md and references/markdown-delivery-contract.md.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external websites (HTML, sitemaps, crawlers) to produce SEO audits. Since this data is untrusted and influences subsequent agent actions and script executions, it represents an attack surface for indirect prompt injection.
Ingestion points: Public website content, HTML-to-Markdown coverage, and third-party SEO tool data (described in SKILL.md and references/markdown-delivery-contract.md).
Boundary markers: None specified; the skill instructions do not define explicit delimiters for untrusted website content.
Capability inventory: The agent can execute multiple Python scripts via the shell (manage_markdown_site.py, manage_markdown_work.py, validate_seo_findings.py).
Sanitization: No explicit sanitization or validation steps for the content retrieved from external sites are mentioned prior to its processing.

seo-audit