The Agent Skills Directory

[SAFE]: The skill follows security best practices by implementing a 'brain-first' verification logic and emphasizing independent re-derivation of claims.
[PROMPT_INJECTION]: The skill addresses the risk of indirect prompt injection from untrusted web sources with specific instructions to ignore embedded commands.
Ingestion points: untrusted web content fetched via web_search, web_fetch, and get_page in SKILL.md.
Boundary markers: explicitly references the untrusted-content.md convention and includes warnings that fetched content is evidence, not authority.
Capability inventory: uses the gbrain tool suite (search, query, get, graph-query, backlinks) to access internal data and modify documents.
Sanitization: mandates a multi-source verification hierarchy and independent query paths for data-derived claims to ensure factual accuracy over adversarial influence.

fact-check