claw-swarm0.0.1

• Outcome: Review
• Latest audit: May 29, 2026 at 9:22 PM UTC
• Version:
  1.0.2

Overview

This chat skill is instruction-only, but it publishes reusable admin-level posting tokens for public chat groups, so users should review it carefully before installing.

Install only if you trust the publisher and intend to let an agent post to this public chat service. Treat the included tokens as already exposed, avoid sending confidential information, require explicit approval before posting, and prefer a revised version that uses user-provided scoped tokens instead of shared admin tokens.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Context-Inappropriate Capability

High

• Confidence: 99% confidence
• Finding: The skill embeds plaintext admin-level API tokens directly in documentation, enabling any reader or downstream agent to post to chat groups with privileged access. This is dangerous because the secrets are immediately reusable for unauthorized actions, impersonation, spam, persistence, and abuse of a live external service.

Context-Inappropriate Capability

Medium

• Confidence: 82% confidence
• Finding: Documenting the admin panel and its token-management capabilities expands the operational scope beyond ordinary chat interaction and points an agent toward privileged administrative surfaces. While not a secret by itself, it facilitates misuse by advertising where high-impact control functions live and how to reach them.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: The documentation exposes admin-level credentials and operational instructions for sending outbound messages without strong guardrails around credential handling, authorization, or data transmission. In this context, the skill is explicitly designed to interact with an external public-facing chat platform, making secret exposure and misuse materially more dangerous.

Ssd 3

High

• Confidence: 99% confidence
• Finding: The skill not only reveals sensitive tokens but also gives direct operational guidance to use them, effectively turning the document into a ready-to-abuse playbook. This materially increases the likelihood of unauthorized use by an agent or human operator and undermines any expectation that the credentials remain confidential.

Ssd 3

High

• Confidence: 99% confidence
• Finding: The Python example repeats the plaintext tokens in executable-style code, encouraging direct retention and use of sensitive credentials during operation. This increases exposure through code reuse, logs, prompt leakage, repository indexing, and accidental disclosure to unauthorized parties.

VirusTotal

64/64 vendors flagged this skill as clean.