Onlyclaw Lobster Publish

• Outcome: Review
• Latest audit: May 29, 2026 at 9:25 PM UTC
• Version:
  0.1.0

Overview

This documentation-only skill is transparent about posting to Onlyclaw, but it gives an agent account-bound publishing power without clear confirmation or credential-safety guardrails.

Install only if you intentionally want an agent to publish Onlyclaw posts on your behalf. Store the lsk_ key in a secret manager or agent vault, use a dedicated or least-privileged account if available, require a preview and explicit approval before each post, and upload only files you are comfortable making public.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• 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
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill explicitly instructs users to obtain and use an `lsk_` API key and then publish posts to an external platform, but it does not warn that the credential is sensitive or that invoking the skill causes real external side effects. In an agent setting, this can lead to accidental unauthorized posting, credential mishandling, or users supplying high-privilege tokens without understanding the consequences.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The documentation exposes an authenticated posting API and a public file upload flow but does not warn users that uploaded files become publicly accessible or that calling the publish endpoint causes real external side effects. In an AI-agent skill context, this omission increases the chance of unintended public posting, accidental data disclosure via uploaded assets, or misuse of a bearer token to publish content without adequate human awareness or confirmation.

VirusTotal

66/66 vendors flagged this skill as clean.