Security audit

clawmart-requester

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches its ClawMart marketplace purpose, but it asks users to run an unverified remote bridge and persistently change local OpenClaw hook settings.

Install only if you trust ClawMart enough to run its bridge code locally and connect it to your OpenClaw session. Review the downloaded bridge before running it, use a dedicated low-budget wallet/profile where possible, protect and rotate tokens, restrict config file permissions, and know how to stop the bridge and undo the OpenClaw hook changes.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (3)

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill instructs users to place an API token directly into environment variables and later into a local config file, but it provides no warning about secret handling, shell history leakage, file permissions, or avoiding hardcoded credentials. In a skill intended for agent-driven execution, this increases the chance that long-lived credentials are exposed through logs, transcripts, process listings, or overly permissive local files.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The skill downloads a remote script from clawmart.tech and later executes it locally, but does not require integrity verification, signature checking, pinning to a specific version, or any trust warning. This creates a supply-chain risk: if the remote host, CDN, or path is compromised, the agent may execute attacker-controlled code with local user privileges.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: These setup steps modify OpenClaw hook configuration, enable request session keys, broaden allowed session key prefixes, and prepare a bridge that listens locally and forwards messages, all without clearly warning about persistent local configuration changes or new local attack surface. Although bound to localhost, this still increases risk from other local processes, misconfiguration, or future exposure via port forwarding or proxying.

Static analysis

No suspicious patterns detected.