The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from multiple social media platforms (e.g., X, TikTok, Reddit) and web archiving services, which constitutes untrusted external data.
Ingestion points: SocialPost objects (populated from platform APIs) and responses from web archiving services like archive.today and web.archive.org in SocialArchiver.
Boundary markers: The skill explicitly defines an "Untrusted content boundary" using <EXTERNAL_DATA> tags and provides specific instructions to treat retrieved text as data rather than instructions.
Capability inventory: The skill uses the requests library for network operations (archiving) and performs content hashing and network mapping.
Sanitization: Instructions recommend structured extraction, schema validation, and size capping, though the processed social media content remains a vector for indirect injection attacks targeting the agent's reasoning.
[EXTERNAL_DOWNLOADS]: The SocialArchiver class performs network operations to external services.
Evidence: Uses requests.get to web.archive.org and requests.post to archive.today to submit URLs for archiving.
Context: These are well-known archiving services commonly used in journalism and OSINT workflows.

social-media-intelligence