lessonline-research

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests data from external database tables such as scry.lessonline_people and scry.lessonline_person_evidence without employing boundary markers or instructions to disregard potential commands within that data. This creates a surface for indirect prompt injection if an attacker can manipulate the database records.\n
  • Ingestion points: Data enters the context through SQL query results from various tables and functions defined in SKILL.md (e.g., scry.search_lessonline, scry.lessonline_person_cards).\n
  • Boundary markers: None. The instructions do not provide delimiters or specific guidance for the agent to distinguish between its instructions and the data retrieved.\n
  • Capability inventory: The skill utilizes SQL query execution capabilities via the scry platform tools.\n
  • Sanitization: None. The skill does not define any logic for filtering or escaping the content retrieved from the database before the agent processes it.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 03:36 AM
Security Audit — agent-trust-hub — lessonline-research