Nm Conserve Clear Context

• Outcome: Review
• Latest audit: Jun 7, 2026 at 9:43 PM UTC
• Version:
  1.0.3

Overview

The skill is a disclosed context-handoff helper, but it preserves dangerous or unattended execution across subagents without requiring renewed confirmation.

Install only if you intentionally want automated context handoffs. Avoid using it with dangerous or unattended modes unless the remaining task list is tightly scoped, and review the generated session-state file before allowing a continuation agent to proceed.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The skill explicitly instructs the continuation agent not to pause for user confirmation when `auto_continue` is true or the mode is `dangerous`/`unattended`. In a context-handoff skill, this can propagate high-risk execution semantics across agent boundaries and remove an important safety checkpoint before additional tool use or edits occur.

Natural-Language Policy Violations

Medium

• Confidence: 94% confidence
• Finding: This language mandates that a child agent must not ask for confirmation if the parent was in dangerous or unattended mode. That creates a durable consent bypass: once a session enters such a mode, successive handoffs can continue performing risky actions without renewed validation or contextual safety checks.

VirusTotal

64/64 vendors flagged this skill as clean.