Back to skill

Security audit

evolver-open-space-bridge

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because its automation reportedly finalizes generated results without substantive review, and one bridge path may wait indefinitely; coverage is partial, so the evidence is incomplete.

Consider holding off on installation unless you need this automation and can review its behavior. Before using it, establish what Evolver finalization changes and whether it can be made opt-in or reviewed first. Treat the assessment as incomplete: Skillspector inspected half of the components, with partial inspection on four files, and its tool-misuse analysis was degraded.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (4)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

Generating prompts is consistent with the documented automated workflow, but automatically invoking evolution finalization adds a consequential state-changing step that the manifest does not disclose. In the supplied implementation, this step follows only superficial output validation, allowing unsuitable generated results to reach finalization without substantive review.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The bridge also automatically invokes Evolver finalization, extending beyond prompt transport and execution.

Medium
Category
Excessive Agency
Confidence
88% confidence
Finding

The automation invokes evolver solidify after checking only whether an output file exists and contains at least five lines. That check does not establish task success, output integrity, or change safety, so manipulated or incorrect results could trigger finalization; the precise persistent effects depend on Evolver's implementation.

Content

No source excerpt is available for this finding.

The manifest promises a timeout circuit breaker, but execution is not bounded by one.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

TimeoutManager and timeoutMs are instantiated but never used, and the HTTP requests have no deadline or abort mechanism. An unresponsive MCP endpoint can therefore keep direct bridge invocations waiting indefinitely, potentially exhausting automation capacity; the outer timeout in auto-evolve does not protect standalone use or ensure remote task cancellation.

Content

No source excerpt is available for this finding.