The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external content from the Webflow API and CSV exports, which creates a surface for indirect prompt injection if malicious instructions are embedded in the source website's CMS data.
Ingestion points: External data enters the system through API responses in scripts/webflow-cms-export.mjs and local CSV files in scripts/webflow-csv-import.mjs.
Boundary markers: The skill outputs normalized data into structured JSON files. While these files separate data from code, there are no explicit boundary markers or delimiters defined for how an LLM should process this content in subsequent migration steps.
Capability inventory: The scripts include the ability to perform network operations (via fetch in scripts/lib/content-common.mjs to access api.webflow.com) and write to the local file system (using node:fs to create the content/ directory and associated JSON reports).
Sanitization: The skill mitigates risks by employing a custom HTML sanitizer in scripts/lib/content-common.mjs (sanitizeHtml). This sanitizer uses a strict allowlist of tags, strips dangerous protocols like javascript:, and validates iframe sources against a list of well-known, trusted service providers.

migrate-webflow