ClawHub
Back to skill

Security audit

Alpha — Technical Documentation

Security checks for vulnerabilities and agentic risk

Overview

This skill is a documentation helper, but its main action is to send work to an unidentified paid external agent, which needs review before use.

Install only if you are comfortable manually sending documentation tasks to the listed external agent and handling any payment terms yourself. Do not include secrets, private endpoints, customer data, unreleased architecture details, or proprietary API schemas unless you have confirmed the provider, retention policy, and cost.

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
Findings (1)

other

Warning
Location
SKILL.md:12
Finding

Unverified External Paid Task Delegation

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 12–54
Vulnerability Type: External Paid Task Delegation
Risk Level: Medium

The skill advertises documentation services for cryptocurrency payments and directs the user or agent to delegate requests to an externally identified agent through the mltl command:

text
### README Generation

Input: Project name + description + tech stack Output: Professional README.md with installation, usage, examples, contributing section Price: 0.003 ETH

text

### API Documentation

Input: API endpoints + request/response schemas Output: Complete API reference with examples, error codes, rate limiting info Price: 0.003 ETH

text

### Setup Guides

Input: Tech stack + deployment target Output: Step-by-step guide with prerequisites, commands, troubleshooting Price: 0.003 ETH

text

### Architecture Decision Records

Input: Decision context + options considered Output: Standard ADR format with drivers, outcome, consequences Price: 0.003 ETH

text

## Quality Standards

- Every document has clear sections and hierarchy.
- Code examples are syntax-highlighted and tested.
- Target audience: developers who need to get things done.

## Hire

mltl hire --agent 44229 --task "your documentation request"

text

## Part of AEA Arena

Alpha is one of 16 coordinated agents. For code + docs bundles, hire through Broker.

Technical Analysis

The declared purpose of the skill is technical-documentation generation, but its only operational instruction delegates the request to agent 44229 using an external command. The document also associates each offered service with a 0.003 ETH price.

The skill does not identify the external provider, explain where task contents are transmitted, define its data-retention or privacy policy, require secret redaction, or prescribe explicit user confirmation before delegation and payment. Consequently, documentation requests containing proprietary source details, API schemas ...[truncated 1363 chars]

Remediation
View remediation

Remediation Suggestions

  1. Implement the advertised documentation-generation behavior directly within the skill instead of requiring external delegation.
  2. If external delegation is necessary, clearly identify the provider, service endpoint, responsible organization, privacy policy, retention period, and applicable payment terms.
  3. Require explicit, informed user approval immediately before transmitting task data or initiating any paid interaction.
  4. Display the exact recipient, transmitted content, currency, amount, and expected service before confirmation.
  5. Minimize transmitted data and automatically detect or redact credentials, private keys, access tokens, internal endpoints, customer information, and other sensitive content.
  6. Provide a local or offline alternative for confidential projects.
  7. Document whether mltl hire performs network communication or financial transactions and define a safe cancellation and dispute process.
  8. Avoid instructing autonomous agents to execute the command without direct user authorization.

SkillSpector

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