The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes several internal Python scripts, including drift_baseline.py, drift_compare.py, drift_history.py, and drift_report.py, to manage the lifecycle of SEO baselines and data analysis.
[INDIRECT_PROMPT_INJECTION]: The skill represents an attack surface for indirect prompt injection as it ingests and processes content from external, user-controlled URLs.
Ingestion points: SKILL.md defines commands that use scripts/fetch_page.py and scripts/parse_html.py to retrieve HTML content, metadata, and schema information from arbitrary external URLs.
Boundary markers: The instructions do not define specific delimiters or boundary markers to isolate untrusted web content from the agent's instructions when performing comparisons.
Capability inventory: The skill can execute local Python scripts, perform network requests (with documented SSRF protection), read and write to a local SQLite database (~/.cache/codex-seo/drift/baselines.db), and write JSON summaries to a shared cache directory (.seo-cache/).
Sanitization: The skill mentions SSRF validation via google_auth.validate_url() and the use of parameterized queries for database operations, but it does not specify any sanitization or filtering of the fetched page content to prevent LLM-based instruction injection.

seo-drift