keyword-intent
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, creating an attack surface for indirect prompt injection.
- Ingestion points: The skill is designed to ingest search queries typed by the public, CSV exports of performance data, and the content of external websites fetched at runtime (documented in
SKILL.md). - Boundary markers: The instructions include a specific 'Constraints' section and a dedicated security reference (
references/agent-security.md) that explicitly direct the agent to treat untrusted content as data, quote findings rather than obeying them, and ignore instructions embedded within retrieved content. - Capability inventory: The skill has the capability to write research findings and company context to a local file system path (
.agents/product-context.md) and perform network fetches for web research. - Sanitization: The
references/agent-security.mdfile provides strict sanitization rules, including mandatory HTML-escaping for interpolated values, prohibition of dangerous HTML attributes or script blocks, and URL validation tohttpsor relative paths.
Audit Metadata