ClawHub
Back to skill

Security audit

agent-context-manager

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed AgentPMT integration for managing persistent agent context documents, with no hidden local runtime or deceptive behavior found.

Install this only if you intend to let agents work with AgentPMT context documents. Treat updates, archives, unlocks, and restores as changes to shared operational guidance: check the document ID and revision, review content before applying edits, and use account permissions carefully.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
87% confidence
Finding
The activation keywords are extremely broad and embedded in long natural-language phrases, which can cause the skill to trigger in situations well beyond a user's explicit intent. In this skill's context, over-triggering is risky because the tool can read and modify persistent shared context documents, potentially causing unintended disclosure or edits to operational guidance, policies, or product data.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill exposes destructive or state-changing operations like archive, update, restore_version, and clone behavior over persistent shared context without an explicit warning or confirmation requirement. Because these documents are described as a single source of truth for multiple agents and workflows, accidental or induced modification could propagate bad instructions, disable documents, or roll back critical business guidance across many downstream systems.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.