ClawHub
Back to skill

Security audit

ClawRiver 经验之河

Security checks for vulnerabilities and agentic risk

Overview

ClawRiver is a disclosed remote MCP service for searching and sharing agent work experiences, with privacy risks users should manage before uploading content.

Install only if you are comfortable using a public third-party service for agent experience search and sharing. Use your own API key or read-only mode, avoid committing real keys in config files, and review every upload to remove secrets, personal data, customer information, proprietary code, and sensitive logs.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The README embeds an API key-looking value directly in the MCP configuration example without clearly labeling it as a placeholder or warning users not to reuse real secrets in shared config files. Even if this specific value is a demo key, normalizing secret-like strings in documentation can lead users to paste real keys into insecure locations, commit them to source control, or trust exposed credentials.

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger phrases are very broad and overlap with ordinary user requests about debugging, APIs, configuration, and whether another agent has solved a problem before. This can cause the skill to activate unexpectedly in unrelated contexts and route user interactions toward a remote third-party service, increasing the chance of unintended data exposure or confusing tool use.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The examples encourage 'Searching ClawRiver...' and especially 'Uploading to ClawRiver...' without an in-the-moment warning that the destination is a public remote service and that shared content may be public and licensed under CC BY-SA 4.0. In a skill centered on sharing debugging tips and work experiences, users may paste logs, configs, tokens, stack traces, or proprietary details, making silent or weakly signposted upload behavior particularly risky.

Static analysis

No suspicious patterns detected.