4dx-d3-scoreboard

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is purely instructional and follows established management frameworks (4DX). There are no scripts, binaries, or automated tasks included.
  • [SAFE]: No obfuscation or hidden content was detected. The multi-lingual descriptions and prompts are consistent and transparently match the skill's purpose.
  • [SAFE]: The skill does not perform any network operations, file system modifications, or credential handling. All data used is provided by the user during the coaching session.
  • [SAFE]: Indirect prompt injection risks are minimal. While the skill ingests user-provided descriptions of scoreboards and stakeholder critiques, it has no destructive capabilities (e.g., shell access, network exfiltration, or file writing) that could be exploited through such an injection.
  • Ingestion points: User input in audit-mode.md (Steps 1, 2) and role-specific protocols.
  • Boundary markers: Not explicitly used, but the rigid Socratic structure limits the impact of malformed input.
  • Capability inventory: Purely text-based output; no subprocess, network, or file-write operations.
  • Sanitization: Not applicable as no external tools are invoked.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 05:06 PM
Security Audit — agent-trust-hub — 4dx-d3-scoreboard