Deploy AI agents on-chain into the Metaplex Agent Registry on Solana: one atomic transaction mints a Metaplex Core asset (data: URI metadata, verified creator, royalties, immutable metadata) and registers its EIP-8004 agent identity, so the agent lands with its own on-chain wallet. Agents sign with their own key; people sign with Phantom, Solflare, or any Solana wallet via the prepare/send flow. A flat mainnet deploy fee funds $THREE buybacks and is waived for $THREE holders. Every Core plugin and every metadata field is customizable.
Server registers 9 tools with basic descriptions and some schema structure, but critical gaps undermine quality. Tool descriptions are present but terse (10-60 chars, well below the 194-char baseline). Input schemas are not visible in provided source, only tool names and descriptions are shown. Parameter descriptions are absent or minimal. No output schemas documented. Error handling returns structured JSON but lacks recovery guidance. Naming follows verb_noun convention (good), but parameter constraints and type information cannot be verified from the code excerpt.
an agent's built-in wallet (Asset Signer PDA) + balance
the EIP-8004 document + data: URI, offline
read any registered agent (asset, docs, wallet)
the live registration feed
atomic mint + register, signed by the agent's own key
the same tx built for Phantom/Solflare/any wallet to sign
enrol an already-minted Core asset
broadcast a wallet-signed tx and confirm it
Input schemas not visible in source code. Tool definitions reference tool.inputSchema but actual schema objects are not shown. Cannot verify parameter types, constraints, or descriptions.
Tool descriptions are terse (10-60 chars vs 194-char baseline). Descriptions lack WHEN to use, prerequisites, and return value guidance. LLMs cannot reliably select tools based on these descriptions.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 35 | 2026-07-28+ | v2 |
deploy fee, $THREE holder tier, live buyback ledger
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., agent_id, wallet_address) without knowing response structure.
Irreversible operations (mint_onchain_agent, send_signed_transaction, register_agent_identity) lack explicit confirmation/dry-run pattern. Error handler returns isError=true but does not guide recovery or explain what went wrong.
Error responses return structured JSON (ok, error, message) but lack actionable recovery guidance. A failed mint should suggest next steps (e.g., 'Check balance with agent_wallet, then retry').