Back to skill

Security audit

getskill

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent OpenClaw skill manager, but it can install and update remote skills into the active agent environment with limited provenance and consent safeguards.

Review this carefully before installing. Use it only if you trust the npm package, the getskill.work service or configured endpoint, and the repositories it returns. Prefer pinned versions and reviewed commits, inspect fetched skill contents before enabling them, avoid elevated global installs, and require confirmation before overwriting existing skills or launching installers.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Error
Location
SKILL.md:31
Finding

Unpinned CLI and Unverified Remote Skill Supply Chain

Content
View full analysis
``` 3. 工具执行以下操作: - 克隆 Git 仓库到 `~/.openclaw/skills-cache/<技能名称>` - 将整个技能目录复制到 `~/.openclaw/skills/<本地技能目录名>`(排除 `.git`) - 返回已复制目录列表 ``` ```markdown ### 配置 API 端点 查看当前端点: ```bash getskill config get ``` 设置自定义端点: ```bash getskill config set https://your-custom-api.com ``` ``` ```markdown ## 关键规则 - 在执行 clone/pull 操作前始终验证 Git 安装 - 复制整个技能目录到 `skills/<技能名>`,并排除 `.git` 目录 - 清理时绝不删除 skills 目录内容 - 只移除缓存 - 支持通过环境变量 `GETSKILL_BASE_URL` 使用自定义端点 - 自动处理 API 重定向(最多 5 次) ``` ### Technical Analysis The Skill directs the agent to install the unversioned `@workskills/getskill` package globally. Because no exact version or integrity hash is specified, package resolution can change after this Skill has been reviewed. Package installation can also invoke package lifecycle scripts unless those scripts are explicitly disabled. The installed CLI obtains repository information from an API, clones the selected repository, and copies the repository contents into the active `~/.openclaw/skills` directory. The documented process does not require: - An immutable CLI version or package integrity hash. - Verification of package signatures or provenance. - An allowlist for API, redirect, or Git repository hosts. - HTTPS certificate or redirect destination restrictions beyond normal client behavior. - Pinning cloned content to an approved commit. - Verifi ...[truncated 2991 chars]
Remediation
View remediation
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

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill states that on Windows it will automatically download a Git installer and launch the setup wizard when Git is missing, but it does not require explicit confirmation or present this as a sensitive action. Automatically retrieving and launching an installer from network-based logic expands the trust boundary and can lead to unexpected code execution, social engineering, or supply-chain compromise if the source or redirection path is abused.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill instructs installation and update flows that clone repositories and copy entire skill directories into ~/.openclaw/skills without an explicit warning that existing local contents may be overwritten or changed. Because this skill manages executable/operational agent skills from remote sources, silent modification of local skill state increases the risk of accidental replacement of trusted skills or persistence of malicious updates.

Content

No source excerpt is available for this finding.