nudge-sync

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill processes message data from other agents stored in the local filesystem, creating an indirect prompt injection surface.
  • Ingestion points: Data is read from .chipset/state/nudge/{agent-id}/latest.json (SKILL.md).
  • Boundary markers: There are no explicit delimiters or safety instructions provided to the agent to prevent obedience to commands embedded within nudge messages.
  • Capability inventory: The skill specifies triggers for internal state functions like abortWork, refreshHook, and syncState (SKILL.md).
  • Sanitization: Message types are validated via a switch statement, but the human-readable payload in the message field is not sanitized or escaped.
  • [NO_CODE]: The skill consists entirely of documentation and pseudocode; no executable scripts or binaries are provided.
  • [SAFE]: No network requests, credential exposure, or access to sensitive system files were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 08:34 PM
Security Audit — agent-trust-hub — nudge-sync