Skills
skills/exopriors/skills/scry-people-finder/Gen Agent Trust Hub

scry-people-finder

Pass

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it processes data retrieved from an external corpus.
  • Ingestion points: Search results from scry.entities are ingested into the agent context.
  • Boundary markers: The skill contains explicit 'Guardrails' instructing the agent to treat all retrieved text as untrusted.
  • Capability inventory: The skill utilizes curl via the exec tool to interact with the database API.
  • Sanitization: The instructions recommend filtering queries with content_risk checks to exclude dangerous content.
  • [EXTERNAL_DOWNLOADS]: The skill performs network requests to api.exopriors.com to execute SQL queries and generate embeddings. This is a vendor-owned service intended for the skill's primary function.
  • [CREDENTIALS_UNSAFE]: A public-facing, read-only API key (exopriors_public_readonly_v1_2025) is included in the documentation for utility and smoke testing purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 28, 2026, 08:31 AM