ClawHub
Back to skill

Security audit

Llm Wiki

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed local Markdown wiki maintenance workflow, with some privacy and activation-scope cautions but no artifact-backed malicious behavior.

Install only if you want an agent to maintain a persistent Markdown wiki in a chosen workspace. Before using it on private notes or documents, tell the agent not to use internet_search unless you explicitly approve the exact query, and confirm the wiki root before any init, ingest, lint, Git, or memory action.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
91% confidence
Finding
The skill is primarily framed as a local wiki maintenance workflow over `raw/` and `wiki/`, but it also advertises `internet_search` for supplementary context without tight scope controls. This can cause the agent to send user topics, source-derived queries, or sensitive project context to external services, expanding the data exposure surface beyond the local workspace model users would reasonably expect.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The trigger conditions include broad terms such as 'wiki', '知识库', 'ingest', and '整理资料', which can match many ordinary conversations unrelated to this skill's full file-modifying behavior. Overbroad activation raises the chance that the agent will unexpectedly begin scanning directories, reading files, or proposing persistent wiki operations when the user intended only a casual discussion.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill description emphasizes knowledge organization benefits but does not upfront warn that it will create directories, write Markdown files, update logs, and possibly initialize Git in the user's workspace. This missing disclosure can lead to uninformed consent and unexpected persistent modifications, especially because the skill is designed around ongoing stateful maintenance.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The documentation mentions `internet_search` without explaining that external search can transmit prompts, keywords, or source-derived context outside the local environment. In a skill centered on processing potentially private documents, the absence of a privacy warning makes accidental data leakage more likely.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.