Back to skill

Security audit

ClawdHub Publish Helper

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent publishing helper, but it tells users to run unpinned ClawHub CLI commands for login and publishing, which is risky in a sensitive workflow.

Install only if you are comfortable with a publishing assistant that reads the skill directory you point it at, creates a publishable copy, and may run ClawHub CLI commands. Prefer replacing `npx clawhub@latest` with a pinned, reviewed CLI version before login or publish, and confirm the exact directory and public slug/version before publishing.

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:113
Finding

Mutable npm Package Executed Through Unpinned npx Commands

Content
View full analysis
--version --name "" /absolute/path/to/skill-dir ``` At `SKILL.md:194-198`: ```bash npx clawhub@latest login npx clawhub@latest whoami npx clawhub@latest publish --slug X --version Y . npx clawhub@latest inspect npx clawhub@latest search ``` ### Technical Analysis The Skill repeatedly instructs users or agents to execute `npx clawhub@latest`. The `latest` distribution tag is mutable and does not identify a fixed, previously reviewed package release. Depending on the local npm cache and configuration, `npx` may retrieve the package from the configured npm registry and execute its code immediately. Consequently, reviewing this repository does not establish which implementation will execute later. A compromised package publisher account, malicious future release, registry compromise, or unsafe registry configuration could cause attacker-controlled code to run without any corresponding change to the audited Skill. The commands include authentication and publication operations. The downloaded CLI may therefore execute in a process with access to ClawHub credentials, npm configuration, the source and publishable Skill directories, and other files available to the invoking account. ### Attack Path 1. An attacker compromises the `clawhub` npm package, its publisher account, its release process, or the registry endpoint used by the victim. 2. The attacker publishes a malicious package release and assigns or causes it to receive the `latest` distribution tag. 3. A user or agent follows the instructions in `SKILL.md` and runs an `npx clawhub@latest` command. 4. `npx` downloads or resolves the attacker-controlled mutable r ...[truncated 1261 chars]
Remediation
View remediation
--version --name "" /absolute/path/to/skill-dir ``` 2. Keep the pinned version consistent throughout `SKILL.md`, including the login, identity, publication, inspection, and search examples. 3. Verify the package name, publisher, source repository, and release provenance before updating the pinned version. 4. Prefer installation through a committed lockfile with integrity metadata where the workflow supports it. Use `npm ci` so dependency resolution follows the reviewed lockfile rather than mutable registry tags. 5. In higher-assurance environments, verify package integrity or provenance and use a trusted registry mirror containing approved artifacts. 6. Require explicit user confirmation before the first download or execution of the third-party CLI, not only before publication. 7. Document a controlled upgrade procedure in which each new CLI version is reviewed and tested before changing the pinned version. ]]>
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 (10)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 197)May include surrounding context.

md
- `SKILL.md` — this file

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README advertises generic trigger phrases like "publish this skill" and "prepare for publishing," which are broad enough to overlap with normal conversation and may cause the skill to activate unintentionally. In a workflow that performs sensitive preparation steps for publication, accidental invocation could expose users to unintended file operations or publishing guidance at the wrong time.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding

The skill instructs use of npx clawhub@latest publish, which pulls and executes the latest package version at publish time. This creates a supply-chain risk: if the upstream package is compromised or a breaking/malicious release is published, the agent or user could execute attacker-controlled code during a high-trust publishing workflow.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding

The changelog example again instructs execution via npx clawhub@latest publish, which dynamically fetches the newest package version. In a publishing workflow that may handle repository contents and credentials, this increases exposure to malicious upstream updates or compromised package releases.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding

The CLI reference lists npx clawhub@latest login, which would execute whatever version is current at the time of use. Because login flows often involve authentication tokens and browser callbacks, running an unpinned package heightens the risk of credential theft or malicious auth interception if the package supply chain is compromised.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding

The whoami reference uses npx clawhub@latest, which still carries the same supply-chain execution risk as other commands. Even though whoami is lower sensitivity than publish, it may still run arbitrary code from the latest package in the user's environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding

The publish reference uses npx clawhub@latest publish --slug X --version Y ., which combines unpinned remote code execution with a sensitive action over local skill contents. If the package were compromised, it could exfiltrate files, modify publish behavior, or steal auth material during publication.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The inspect example uses npx clawhub@latest, which is another instance of executing an unreviewed latest package version. While the action itself is relatively low sensitivity, the execution model still permits arbitrary upstream code to run in the user's environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The search example also relies on npx clawhub@latest, exposing users to the same supply-chain risk from remote package execution. Although searching is less sensitive than publishing or logging in, compromise of the latest package could still lead to arbitrary code execution.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.