find-conferences

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill configuration uses restricted bash commands and scopes file operations to specific local paths. No malicious patterns, obfuscation, or data exfiltration attempts were identified.- [SAFE]: Analysis of the indirect prompt injection surface shows that while the skill ingests untrusted web data via WebSearch and WebFetch, it lacks the dangerous capabilities required for a successful exploit.
  • Ingestion points: External web content retrieved via WebSearch and WebFetch (Step 2).
  • Boundary markers: None explicitly defined to separate untrusted web content from system instructions.
  • Capability inventory: Restricted Bash tools (mkdir, test, date), Read (ground-truth.md), and Write (conferences.md).
  • Sanitization: No programmatic sanitization is specified for retrieved web content.
  • Risk Assessment: The risk is negligible as the agent's capabilities are limited to formatting information into Markdown files and performing basic directory management.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 05:16 AM
Security Audit — agent-trust-hub — find-conferences