c2

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by interpolating user-provided research questions and constraints into its prompts without explicit delimiters or sanitization instructions.
  • Ingestion points: Untrusted data enters via the {research_question}, {resources}, and {constraints} placeholders in the prompt template within SKILL.md.
  • Boundary markers: Absent; the user input is directly embedded into the expert instructions.
  • Capability inventory: The skill uses tools/commands to read configuration files (config/diverga-config.json) and update decision logs (.research/decision-log.yaml).
  • Sanitization: No evidence of input validation or escaping for the interpolated user content.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 10:10 AM
Security Audit — agent-trust-hub — c2