Use Cases & Examples
1. DAO Voting Assistant Agent
Scenario: A governance bot powered by an LLM helps a user understand and vote on DAO proposals based on their on-chain identity, DAO membership, and past votes.
Flow:
User connects wallet (0xabc...)
Frontend requests context with scopes:
["read.dao", "read.tokens", "read.profile"]Backend retrieves MCO via GET /context
Agent receives context and generates a voting suggestion
Result: LLM outputs a summary and voting suggestion based on user history and voting power.
2. zk-KYC AI Support Bot
Scenario: A user seeks access to age-restricted content or services. Instead of exposing personal info, they submit a zkProof via MCP3 context.
Flow:
User generates ZK context using a Semaphore circuit
LLM verifies proof via POST /verify
Assistant confirms eligibility without leaking personal data
3. Contextualized AI Wallet Companion
Scenario: A browser plugin AI agent advises the user on token usage, NFT trends, and portfolio stats.
Flow:
Plugin refreshes MCO every 60s with relevant scopes
LLM prompt includes wallet and NFT data
Agent gives personalized advice
4. Reputation-based AI Filtering
Scenario: An LLM API uses MCP context to modulate input filtering based on social reputation (e.g. Gitcoin Passport score, Farcaster account age, DAO rank).
Flow:
Context payload includes social and attestation data
LLM adjusts moderation threshold dynamically
5. Delegated Agent Using Stored Context
Scenario: A user stores their signed context object in IPFS or Arweave and lets agents act on their behalf asynchronously.
Flow:
Signed MCO stored at
ipfs://QmXYZ...Delegate agent fetches, verifies, and uses it for actions (DAO, airdrops, auto-voting)
Advanced Patterns
Context Chaining
Agent combines multiple MCOs over time to infer long-term intent.
Multi-Wallet Context
User owns several wallets; MCO merges identities.
ZK-Only Context
No raw data, all claims proven via ZK circuits.
Agent-to-Agent Context Passing
One AI agent creates and signs context for another (e.g., LLM-to-LLM).
Last updated