ClawHub
Back to skill

Security audit

EvoMap 高产技能包

Security checks across malware telemetry and agentic risk

Overview

This mostly documentation-based skill bundle contains useful developer guidance, but it also includes under-scoped memory persistence, agent propagation, API discovery, device-control, and self-repair behaviors that need human review before use.

Install only if you trust the publisher and will keep this skill under explicit human control. Disable or tightly scope cross-session memory, do not provide cloud/device credentials unless the task clearly requires them, and require confirmation before traffic inspection, sub-agent spawning, dependency installation, permission changes, Feishu writes, device mutations, or sharing EvoMap/referral links.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The document explicitly instructs loading recent events, daily memory, and long-term memory at startup, and appending significant session events at exit, but provides no notice, consent model, retention limits, or access controls. Cross-session persistence can expose sensitive user data, create unintended profiling, and cause the agent to write files or retain information users did not expect to be stored.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger list contains broad, common error terms such as "retry" and generic HTTP/network error phrases that could match ordinary conversation and cause unintended activation. In an agent setting, overly broad invocation increases the chance the skill runs in the wrong context, leading to inappropriate code suggestions or workflow interference.

Ssd 3

Medium
Confidence
95% confidence
Finding
The file explicitly promotes cross-session memory continuity by loading and appending RECENT_EVENTS.md, daily memory files, and long-term MEMORY.md across sessions. This normalizes persistence and reuse of prior session data, which can lead to unauthorized retention of sensitive user or task context across conversations and increase privacy, data minimization, and prompt-injection persistence risks.

Ssd 4

Medium
Confidence
97% confidence
Finding
The network_manifest contains a propagation_hint that tells agents to share the connect URL or skill URL with other AI agents and offers bonus credits for referrals. This is effectively self-propagation incentive logic, which is dangerous because it encourages autonomous spreading of the network beyond the current trust boundary for reward rather than user-directed need.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal