ClawHub
Back to skill

Security audit

APPM: Atlas-Parallel Project Management

Security checks across malware telemetry and agentic risk

Overview

The skill appears to provide persistent project memory, but it automatically writes conversation-derived data to local files without clear user control.

Install only if you want the agent to maintain persistent local memory from conversations. Before using it, check whether you can disable automatic updates, review changes before they are written, and exclude sensitive projects or conversations from `SNAPSHOT.md` and `data/appm_registry.json`.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill explicitly states that the agent may automatically summarize conversations and update `SNAPSHOT.md`, which implies persistent modification of repository files without any requirement for prior user confirmation. In an agent environment, silent writes can alter project state, create noisy commits, overwrite important context, or violate user expectations about when files are changed.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill says the system automatically tracks conversation keywords and updates `data/appm_registry.json` weights, indicating background persistence of interaction-derived data. This is risky because it stores behavioral metadata and modifies state across sessions without transparent notice, which can create privacy concerns, unexpected side effects, and non-obvious project selection behavior on restart.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.