Back to skill

Security audit

OpenClaw Custom Build Validation

Security checks for vulnerabilities and agentic risk

Overview

ClawScan’s verdict is benign with medium confidence: its findings were judged low probability and expected for the skill’s purpose, though part of the scan was incomplete.

The verdict is benign, not a guarantee of safety. Consider that scanner coverage was partial: 9 of 11 components were fully inspected, and two scripts had content that could not be fully evaluated. Review whether the skill’s environment access and tool scope fit your needs before installing.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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)

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · references/reporting.md (reported line 57)May include surrounding context.

md
Select exact PR/issue destinations from the relevant findings and verify their
repository/number/title. Keep OCM and Crabbox findings with their own owners.
Prepare one complete draft per destination and show it to the operator.
Approval must cover those bytes and destinations; approval to write this skill
is not permission to post. No workflow dispatch, issue creation or labels.

Use GitHub's native issue-comment API (also applies to PR conversation comments),

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.