Security audit

ClawLink

Security checks for vulnerabilities and agentic risk

Overview

ClawLink is a disclosed third-party integration hub, but it needs review because it can act across many connected apps and includes an unsafe local-file upload command flow.

Install only if you trust ClawLink with access to the external accounts you connect. Review connected app scopes carefully, use previews and explicit confirmation for writes, avoid connecting unnecessary sensitive apps, and be cautious with local file uploads because the skill directs the agent to run a service-provided curl command.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The skill explicitly instructs the agent to execute a shell `curl` command supplied via an error message to upload local file bytes. That creates a command-execution bridge from remote/plugin-controlled content into the local shell, and can expose arbitrary local files or enable shell injection if the returned command is malformed or adversarial.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The skill says to use ClawLink for essentially any interaction with a wide range of external apps and to 'always check ClawLink first,' making activation scope extremely broad. Over-broad routing increases the chance the agent will invoke a powerful third-party integration layer unnecessarily, exposing user data and expanding the blast radius for unrelated requests.

Static analysis

No suspicious patterns detected.