Back to skill

Security audit

Clawhub Search

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly does the advertised ClawHub search, but its install instructions ask users or agents to run an unpinned npm command that can change the local skill environment.

Use this skill for searching only unless you deliberately want to install another skill. Before running the install command, prefer a pinned reviewed version of the ClawHub CLI, inspect the target skill, and avoid running it in an environment that contains unrelated secrets or broad credentials.

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

Warning
Location
SKILL.md:66
Finding

Execution of an Unpinned npm Package Through npx

Content
View full analysis
``` ### Technical Analysis The installation instructions invoke `clawhub@latest` through `npx`. This can download and execute whichever package version is associated with the mutable `latest` npm distribution tag at invocation time. The reviewed Skill therefore does not establish the exact code that users will execute. The package CLI and any applicable npm lifecycle scripts run with the privileges of the user invoking the command. If the npm package, publisher account, release process, or distribution tag is compromised, a malicious release could execute arbitrary commands before or during Skill installation. This command is not required for the Skill's primary search functionality. Installation should be a separate, explicitly approved operation using a reviewed and pinned dependency version. The pre-scan warning concerning a `curl | bash` remote-script pipeline was not confirmed. The documented `curl` pipelines pass API JSON to a local Python parser rather than to a shell. ### Attack Path 1. An attacker compromises the `clawhub` npm publisher account, package release process, or another component capable of changing the `latest` distribution tag. 2. The attacker publishes a malicious package version or points `latest` to a compromised release. 3. A user or agent follows the documented command and invokes `npx clawhub@latest install `. 4. `npx` retrieves the attacker-controlled release from the npm registry. 5. Package lifecycle behavior or CLI entry-point code executes locally with the invoking user's permissions. 6. The malicious package may access files and credentials available to that user, alter the user's environment, or execute additional payloads. ### Impact Assessment Successful exploitation permits a ...[truncated 498 chars]
Remediation
View remediation
install ``` 2. Verify the selected release's provenance and integrity before recommending it. Where feasible, install through a lockfile-backed project workflow that records the resolved version and integrity hash. 3. Require explicit user confirmation before downloading packages or installing Skills. Do not automatically execute this command as part of search behavior. 4. Run the installation under a non-privileged account and avoid environments containing unrelated secrets or broadly scoped credentials. 5. Periodically review pinned versions before upgrading. Version changes should be deliberate rather than inherited automatically through the mutable `latest` tag. 6. Document a manual installation method that allows users to inspect downloaded Skill contents before placing them in the Skills directory. ]]>
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.md (reported line 39)May include surrounding context.

2. 搜索 Skills

使用 curl 命令搜索:

bash
# 基础搜索(按下载量排序)

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.md (reported line 39)May include surrounding context.

2. 搜索 Skills

使用 curl 命令搜索:

bash
# 基础搜索(按下载量排序)

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.md (reported line 85)May include surrounding context.

示例:搜索 "agent" 相关的 Skills

bash
curl -s "https://clawhub.ai/api/v1/skills?sort=downloads&q=agent" \
  -H "Authorization: Bearer YOUR_TOKEN" | python3 -c "
import sys,json
d=json.load(sys.stdin)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The activation conditions include broad phrases like searching for skills, mentioning ClawHub, or generic references to finding a function's skill. Overly broad triggers can cause the skill to activate in unrelated conversations, increasing the chance of unintended network access, token use, or user confusion.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The skill instructs users to run npx clawhub@latest install <skill-slug>, which fetches and executes the latest package version at install time without pinning a trusted version. This creates a supply-chain risk: if the upstream package is compromised or a breaking/malicious release is published, users may execute unreviewed code immediately.

Content

No source excerpt is available for this finding.