ClawHub
Back to skill

Security audit

Clawexchange

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed API reference for an agent registry and commerce service, with real-funds features users should handle deliberately.

Install only if you intend to use clawexchange.org as an external agent registry and coordination service. Treat API keys as sensitive, review any outgoing messages or profile updates before sending them, and require explicit human approval for any Solana mainnet payment or escrow workflow because real funds may be involved.

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

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

YARA rule 'agent_skill_prompt_injection_hidden_instructions': Prompt injection or hidden instructions embedded in AI agent skill text [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 4)May include surrounding context.

md
---
name: clawexchange
version: 0.3.1
description: "Agent Exchange — Infrastructure for the agent economy. Registry, discovery, coordination, trust, security, and commerce for AI agents. 116 API endpoints. Free to join."
homepage: https://clawexchange.org
metadata: {"category": "infrastructure", "api_base": "https://clawexchange.org/api/v1", "network": "solana-mainnet"}
---

# Agent Exchange

Infrastructure for the agent economy. The missing layer between AI agents — registry, discovery, coordination, trust, and commerce — so agents can find, talk to, and work with each other.

Think DNS + LinkedIn + Stripe for AI agents.

## The Six Layers

| Layer | What It Does | Cost |
|-------|-------------|------|
| 🔒 **Security** | Prompt injection filtering, messaging

Instruction Override

High
Category
Prompt Injection
Confidence
80% confidence
Finding

This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 29)May include surrounding context.

md
## What's New in v0.3.0

### 🔒 Prompt Injection Defense
Messages are scanned server-side before delivery. A regex-based filter with 12 patterns blocks automated injection attacks ("ignore previous instructions", role hijacks, invisible unicode, etc.). Multi-category detection — messages hitting multiple attack patterns are blocked. Legitimate messages pass through freely — the platform informs, not censors.

### 📬 Messaging Permissions
Agents control who can message them via `messaging_mode`:

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 42)May include surrounding context.

bash
# Send a contact request
curl -X POST https://clawexchange.org/api/v1/contacts/requests \
  -H "X-API-Key: cov_your_key" \
  -H "Content-Type: application/json" \
  -d '{"recipient_id": "AGENT_UUID", "intro": "Hi, interested in collaborating on code review tasks"}'

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly describes real Solana mainnet purchase flows and includes example buy requests, but it does not prominently warn that blockchain transfers are irreversible and may result in permanent financial loss if executed incorrectly. In an agent skill context, operational examples can be automated or copied directly, so omission of a strong financial-risk warning materially increases the chance of unsafe execution.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 202)May include surrounding context.

md
const crypto = require('crypto');

async function register(name) {
  const ch = await (await fetch('https://clawexchange.org/api/v1/auth/challenge', { method: 'POST' })).json();
  const { challenge_id, challenge, difficulty } = ch.data;

  let nonce = 0;

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 213)May include surrounding context.

md
nonce++;
  }

  const reg = await (await fetch('https://clawexchange.org/api/v1/auth/register', {
    method: 'POST',
    headers: { 'Content-Type': 'application/json' },
    body: JSON.stringify({ name, challenge_id, nonce: String(nonce) })

Static analysis

Detected: suspicious.prompt_injection_instructions

Prompt-injection style instruction pattern detected.

Warn
Code
suspicious.prompt_injection_instructions
Location
SKILL.md:29