ClawHub

Circleci

Security checks across malware telemetry and agentic risk

Overview

This is a real CircleCI integration, but it gives agents broad authenticated CI/CD control without clear safeguards before changes or deletes.

Install only if you trust Membrane and intend to let the agent operate your CircleCI account. Use the least-privilege CircleCI connection available, limit it to the intended organization or project where possible, and require explicit approval before changing environment variables, triggering or canceling workflows, deleting contexts, or using raw proxy API requests.

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 (2)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill documents destructive actions such as rerunning, canceling, and deleting CircleCI resources without any guidance to require explicit user confirmation before execution. In an agent setting, this increases the risk that a model could perform state-changing or disruptive operations based on ambiguous prompts or overbroad automation.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The proxy-request section enables arbitrary authenticated API calls, including destructive HTTP methods like POST, PATCH, and DELETE, but provides no warning that these requests can modify remote resources or transmit sensitive data over the network. This broad escape hatch bypasses the safety of curated actions and can let an agent perform unintended high-impact operations against CircleCI.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal