ClawHub
Back to skill

Security audit

nvimclaw

Security checks for vulnerabilities and agentic risk

Overview

This skill gives an agent controlled access to a live Neovim session, and its sensitive editor access is disclosed and aligned with that purpose.

Install only if you want OpenClaw agents to inspect and potentially edit your live Neovim buffers. Keep privileged tools disabled unless needed, prefer explicit command allowlists on shared gateways, avoid sending chat turns from sensitive buffers unless the automatic context attachment is acceptable or disabled, and use confirmation for destructive Ex commands.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The send-from-Neovim section states that a <nvimclaw_context> envelope is auto-attached with path, cursor, changedtick, and a 41-line content window, but it presents this as behavior rather than as a privacy/security warning. Because messages are sent into a persistent agent session with shared memory/persona, this can unintentionally disclose sensitive source code, secrets, or nearby text from the user's editor context.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill prominently advertises read/edit access to a live Neovim instance, including unnamed buffers, but the opening description does not explicitly foreground that privileged commands can immediately mutate the user's active in-memory editor state, not just files on disk. In this context, that can lead users or downstream agents to underestimate the risk of destructive edits to unsaved buffers, active windows, or the wrong target buffer.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 202)May include surrounding context.

Full-buffer overwrite. Alias for replace_lines(0, -1, lines) with the same conflict semantics. Provided for agents trained on vscode.file.write; prefer nvim.ex.substitute or nvim.buffer.replace_lines when possible — they preserve Vim's undo history per edit.

bash
openclaw nodes invoke --node <N> --command nvim.buffer.write \
  --params '{
    "path": "drafts/example.md",
    "content": "Schopenhauer is hilarious. ...",

Static analysis

No suspicious patterns detected.