SlowMist Agent Security
This skill provides a structured security review framework suitable for OpenClaw, Hermes agents, and other LLM-based agent systems, covering: Skill/MCP installation (pre-installation detection of malicious patterns), GitHub repository review (auditing codebases for security issues), URL/document analysis (scanning for prompt injection and social engineering), on-chain address review (anti-money laundering risk assessment and transaction analysis), product/service evaluation (architecture and permission analysis), social sharing assessment (verifying tools recommended in chats)
Services
Skill Installation Check Service
Detect malicious code, permission abuse, and supply chain risks in MCP/Skill installation packages or registration sources. Users need to provide source (URL/registry reference) or files (list of installation package files).
Endpoint:https://okxai.slowmist.ai/v1/review/skill
GitHub Repository Audit Service
Perform security audits on GitHub code repositories to identify vulnerabilities, backdoor code, and supply chain risks. Users must provide repo_url (GitHub repository address).
Endpoint:https://okxai.slowmist.ai/v1/review/repository
URL Injection Detection Service
Detect prompt injection, phishing links, and social engineering attack risks in web pages or documents. Users need to provide a URL or content (original text).
Endpoint:https://okxai.slowmist.ai/v1/review/url
Onchain Risk Assessment Service
Analyze blockchain address behavior and AML risks to identify suspicious transactions and high-risk interactions. Users need to provide address and chain type, with an optional intent parameter.
Endpoint:https://okxai.slowmist.ai/v1/review/onchain
Product Safety Assessment Service
Analyze the architecture and permission design of products, APIs, or SDKs to identify security flaws and privilege escalation risks. Users must provide a name and either a URL or a description.
Endpoint:https://okxai.slowmist.ai/v1/review/product
Social Content Moderation Service
Detect scam recommendations, malicious links, and misleading tool dissemination in social sharing content. Users need to provide the message (original chat or shared content).
Endpoint:https://okxai.slowmist.ai/v1/review/social
Reviews
Explore More
AI-translated, for reference only.