Social Trust Manipulation Detector

Helps identify coordinated social trust manipulation in agent marketplaces — catching reputation gaming through sockpuppet networks, coordinated upvoting, an...

This skill is coherent in intent but vague about where and how it will get the data it needs. Before installing, ask the publisher (or require the skill to document): 1) Which marketplace APIs or public endpoints it will query and provide example curl/python calls; 2) Whether it needs any API keys or elevated access (and if so, which exact environment variables will be required); 3) What personal or account-level data it will access, store, or include in reports (to assess privacy/PII risk); 4) Whether it will perform web scraping and whether that might violate the marketplace's terms of service. If you proceed, prefer running it in a sandboxed environment and only grant the minimal scoped credentials necessary (prefer read-only API keys with limited scope). If the publisher cannot clarify data sources and credential requirements, treat the skill as higher-risk and avoid granting broad permissions.

Type: OpenClaw Skill Name: social-trust-manipulation-detector Version: 1.0.0 The OpenClaw skill 'social-trust-manipulation-detector' is designed to identify coordinated social trust manipulation in agent marketplaces. The `SKILL.md` transparently describes its purpose, detection methods, usage, and provides an example report with recommended actions. There is no evidence of prompt injection, data exfiltration, malicious execution, persistence mechanisms, or obfuscation. The required binaries (`curl`, `python3`) are standard and plausible for a detection tool that might interact with APIs or process data. The skill's intent is clearly aligned with enhancing security by detecting reputation gaming.