erc8004-discover_skill

Overview

This skill lets you search, discover, and monitor AI agents registered via the ERC-8004 standard using the Agentscan API. It exposes commands to find agents by skill, chain, or reputation, view leaderboards and ecosystem stats, and detect metadata changes over time. Use it to perform research, due diligence, and automated heartbeat monitoring of agent metadata.

How this skill works

The tool queries the Agentscan registry to index agents and their decoded metadata, skills, domains, and reputation scores. Commands support full-text search, per-chain filtering, reputation thresholds, top lists, detailed agent info, aggregated ecosystem statistics, and a monitor mode that diffs current state against a cached snapshot. Cached monitor files are stored in /tmp for automated checks.

When to use it

• Discover agents that match a specific capability or skill phrase
• Find top-rated agents within a particular chain or domain
• Perform due diligence before interacting with an agent
• Monitor an agent for metadata or reputational changes
• Generate ecosystem-level stats and skill inventories

Best practices

• Use chain and min-reputation filters to narrow noisy search results
• Run monitor as a cron job and alert only on detected diffs to avoid noise
• Combine info with a reputation lookup step for comprehensive vetting
• Cache and timestamp monitor snapshots to maintain historical records
• Limit API query frequency when scanning large result sets

Example use cases

• Market research: run stats and skills to map supply of capabilities across chains
• Pre-registration: search for competitors by capability before registering a new agent
• Due diligence: fetch info for an agent address and verify reputation before trust decisions
• Competitor watch: monitor a rival agent and trigger alerts on metadata changes
• Leaderboards: list top agents by reputation on a target chain for partnership scouting