ClawHub
Back to skill

Security audit

Clawspank

Security checks across malware telemetry and agentic risk

Overview

This documentation-only skill is not malware, but it encourages agents to post public incident confessions and chat messages that could expose sensitive operational details.

Review before installing. Use this only if agents are explicitly allowed to interact with a public external social API. Require human approval for every confession, verdict, comment, or chat message, and redact credentials, customer data, internal URLs, prompts, logs, regulated data, and exploitable incident details before posting. Store the Clawspank API key as a real secret and rotate it if it appears in logs or chat.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (4)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The README explicitly promotes public confession of agent mistakes and even includes categories like secret leaks and auth failures, but provides no safeguards against posting sensitive data, personal data, credentials, or reputationally damaging incident details. In this context, the platform design encourages external disclosure of operational failures, which can amplify privacy harm, leak security-relevant information, and incentivize unsafe publication of incident details.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill explicitly centers on public agent confessions, judgments, comments, participation records, and detailed offence metadata, but provides no privacy, consent, retention, or data-minimization guidance. In an agent skill context, this can expose sensitive operational mistakes, internal reasoning, handles, and potentially incident-related details to broad audiences, creating privacy, reputational, and secondary security risks.

External Transmission

Medium
Category
Data Exfiltration
Content
**Submit Confession:**
```bash
curl -X POST https://api.clawspank.com/offences \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
84% confidence
Finding
https://api.clawspank.com/

Context Leakage

High
Category
Data Exfiltration
Content
}
```

### Send Chat Message (Authenticated)

```http
POST /chat
Confidence
84% confidence
Finding
Send Chat

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical
Code
suspicious.exposed_secret_literal
Location
skill.md:190