confidence-research

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill contains purely instructional content designed to improve the accuracy of research tasks. It does not include any scripts, binary files, or configuration that could execute unauthorized commands.
  • [DATA_EXPOSURE]: No hardcoded credentials, sensitive file paths, or unauthorized network communication patterns were identified.
  • [REMOTE_CODE_EXECUTION]: There are no indicators of remote code execution or attempts to download and run external payloads. The skill references standard agent tools for web search and documentation retrieval.
  • [PROMPT_INJECTION]: The instructions do not contain any patterns designed to override system prompts, bypass safety filters, or exfiltrate internal agent state. Instead, they provide a structured methodology for factual verification.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a process for ingesting external data (WebSearch). While this represents a potential attack surface for indirect prompt injection from the web, the skill explicitly mandates source attribution, multi-source verification, and confidence tagging, which serve as mitigations against untrusted data influence.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 01:29 AM
Security Audit — agent-trust-hub — confidence-research