Beckmann Knowledge Graph × Self-Improving + Proactive Agent

• Outcome: Pass
• Latest audit: May 28, 2026, 4:36 AM
• Version:
  1.0.0

Overview

This is an instruction-only orchestration skill whose memory use and deep-reasoning escalation behavior are disclosed and aligned with its stated purpose.

Before installing, review the two required base skills because this skill mostly orchestrates their behavior. If used, periodically inspect ~/self-improving/ memory files, avoid saving secrets there, and only confirm Beckmann escalation when you want a slower, more speculative deep-reasoning workflow.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Vague Triggers

Medium

• Confidence: 80% confidence
• Finding: The trigger criteria are broad and subjective, including categories like 'strategic dead end,' 'AI safety,' and 'epistemological limit,' which can cause the agent to suggest switching in many normal conversations. While the README says confirmation is required before acting, overbroad proactive triggering can still steer user behavior, create prompt-routing confusion, and unnecessarily expose users to a more speculative reasoning mode.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The escalation triggers are broad and subjective enough that many normal user questions could be routed into the Beckmann workflow unnecessarily. In a meta-skill that changes reasoning mode, this can cause scope creep, inconsistent handling, and over-collection or persistence of user-derived insights into memory files without a clear necessity threshold.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The Complexity Check relies on a highly subjective heuristic that different agents could interpret inconsistently, making escalation behavior unpredictable. That unpredictability is risky here because escalation changes processing flow and may trigger additional memory logging, graph loading, and persistence of analysis artifacts for routine questions.

VirusTotal

VirusTotal findings are pending for this skill version.