Back to skill

Security audit

Chain Sensei

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly consistent with blockchain analysis, but it needs review because it describes automatic paid x402 use and an unpinned remote installer.

Review before installing. Use a pinned, trusted installer version where possible, and do not enable x402 payment capability unless you have explicit spending limits, per-request approval, and a way to disable alerts or recurring triggers.

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:80
Finding

Unpinned Package Execution Through npx

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 80-83
Vulnerability Type: Unpinned third-party installer execution
Risk Level: Medium

Complete Code Snippet:

markdown
**1. Install the skill:**
```bash
npx clawhub@latest install chain-sensei
text

### Technical Analysis

The installation instructions invoke `clawhub` through `npx` using the mutable `@latest` distribution tag. If the package is not already available locally, `npx` can retrieve it from the configured package registry and execute it with the invoking user's permissions.

Because `@latest` does not identify an immutable, reviewed release, the code executed by this command may change after this skill has been audited. The project does not provide an exact dependency version, integrity hash, lockfile, vendored installer, or other mechanism for verifying that the retrieved package matches a trusted artifact.

This is a supply-chain weakness rather than evidence that the current `clawhub` package is malicious. The same unsafe command also appears in `README.md` at lines 5-9, increasing the likelihood that users will follow it.

### Attack Path

1. An attacker compromises the registry account, publication pipeline, or another distribution component used to publish `clawhub`.
2. The attacker publishes a malicious release and assigns it to the mutable `latest` tag.
3. A user follows the documented `npx clawhub@latest install chain-sensei` instruction.
4. `npx` downloads the attacker-controlled package from the configured registry.
5. The retrieved package executes with the permissions and environment available to the user running the installation.
6. The malicious package can perform any operation permitted to that user before or during installation.

### Impact Assessment

Successful exploitation could provide arbitrary code execution under the invoking user's account. Depending on that account's privileges and environment, the d
...[truncated 549 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace @latest with an exact, security-reviewed package version.
  2. Verify the downloaded package using registry integrity metadata or a documented cryptographic checksum.
  3. Record installation dependencies in a lockfile where the surrounding package-management workflow supports it.
  4. Document the expected registry and reject untrusted registry overrides.
  5. Prefer an installation process that separates package download and verification from execution.
  6. Run the installer with a minimally privileged account in a sandbox or isolated environment without unnecessary credentials.
  7. Keep the installation command in README.md and SKILL.md synchronized, applying the same pinned and verified version in both locations.
  8. Review each new installer release before updating the pinned version.
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (5)

Rp1

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

The README instructs users to run npx clawhub@latest install chain-sensei, which pulls and executes the latest published version of the package at install time rather than a reviewed, pinned release. If the package is compromised, a malicious version is published, or a dependency is hijacked, users and agents following the instructions could execute attacker-controlled code during installation.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
87% confidence
Finding

The installation command uses npx clawhub@latest, which fetches and executes the most recent package version at install time rather than a reviewed, fixed version. That creates a supply-chain risk: if the package is compromised or a breaking/malicious update is published, agents may execute untrusted code during installation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The usage guidance tells the agent to 'just ask naturally' with broad prompts like 'What's the risk of 0x1234...?' and 'Is this wallet safe to interact with?' without defining when the skill should activate or what requests are out of scope. In an agent setting, this can cause over-broad invocation, prompt-triggering by untrusted content, or unintended use of the skill in contexts where on-chain analysis was not explicitly requested.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The premium guidance says the agent will 'automatically handle x402 payment negotiation' and only notes that USDC on Base is needed, but it does not prominently frame this as a charge-incurring action requiring explicit consent or spending controls. In autonomous agents, this can lead to unexpected micropayments, repeated charges, or budget drain if premium paths are triggered implicitly.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The API reference states that 'All requests use natural language. The skill interprets and executes,' which implies unrestricted natural-language execution semantics. Without narrowing accepted intents or schema, an integrating agent may hand arbitrary text to the skill, increasing the risk of unintended actions, misrouting, or abuse triggered by adversarial prompt content.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.