brickken agent groups the ERC-8004 identity and reputation operations. Every subcommand is prepare-only by default; add --execute to prepare, sign locally, send, and pay through x402 in one step.
For the field-level semantics of each underlying method, see Agentic Methods with x402.
Shared options
These options apply to every agent subcommand:
| Flag | Description |
|---|
--chain <chain> | Chain identifier (decimal or hex) |
--signer-address <address> | Signer wallet address |
--gas-limit <value> | Optional explicit gas limit |
-f, --file <path> | JSON/YAML input file (merged with CLI flags) |
--execute | Prepare, sign locally, send, and pay through x402 |
--owner-email / --email is optional metadata. Agentic operations do not require a tokenizer user in the Brickken database.
agent register
Prepare or execute an ERC-8004 agent registration → backend method agentRegister. The prepare response includes info.agentUuid; save it for the follow-up set-uri and set-metadata calls.
| Flag | Type | Required | Description |
|---|
--name <name> | string | Yes | Agent display name |
--description <value> | string | Yes | Agent description |
--image <url> | string | Yes | Agent image URL or IPFS URI |
--service-name <name> | string | Yes | Single service name |
--service-endpoint <url> | string | Yes | Single service endpoint |
--service-version <version> | string | No | Single service version |
--ai-model-name <name> | string | No | AI model name |
--ai-model-provider <provider> | string | No | AI model provider |
--tag <value> | string | No | Discovery tag (repeatable) |
--version <value> | string | No | Agent profile version |
--documentation <url> | string | No | Documentation URL |
--source-code <url> | string | No | Source code URL |
--license <value> | string | No | License identifier or URL |
--agent-type <value> | string | No | Agent type |
--supported-trust <value> | string | No | Supported trust capability (repeatable) |
--x402-support <value> | boolean | No | Whether the agent advertises x402 support |
--active <value> | boolean | No | Whether the agent is active |
--owner-email <email> | string | No | Tokenizer owner email |
--email <email> | string | No | Alias for --owner-email |
brickken agent register \
--chain "$CHAIN" \
--signer-address "$WALLET" \
--name "Research Agent" \
--description "On-chain AI research agent" \
--image https://example.com/agent.png \
--service-name A2A \
--service-endpoint https://agent.example/.well-known/agent-card.json \
--service-version 0.3.0 \
--ai-model-provider OpenAI \
--ai-model-name "Research Model" \
--x402-support true \
--execute \
--json | tee register-output.json
export AGENT_UUID="$(jq -r '.prepared.info.agentUuid' register-output.json)"
agent set-uri
Prepare or execute an agent URI/profile update → agentSetURI. Run this after the registration transaction is mined; it injects the real on-chain agentId and publishes the final profile to IPFS.
In addition to the shared options and all register profile flags (--name, --description, --image, --service-*, --ai-model-*, --tag, --version, --documentation, --source-code, --license, --agent-type, --supported-trust, --x402-support, --active):
| Flag | Type | Required | Description |
|---|
--agent-uuid <uuid> | string | Yes¹ | Stored tokenized agent UUID |
--agent-id <id> | string | Yes¹ | On-chain ERC-8004 agent ID |
--agent-uuid or --agent-id.
brickken agent set-uri \
--chain "$CHAIN" \
--signer-address "$WALLET" \
--agent-uuid "$AGENT_UUID" \
--name "Research Agent" \
--description "On-chain AI research agent" \
--tag ai-agent \
--documentation https://docs.brickken.com \
--source-code https://github.com/Brickken/brickken-api-cli \
--license MIT \
--agent-type research \
--supported-trust feedback \
--x402-support true \
--active true \
--execute \
--json
agentSetMetadata. Writes a single key/value pair on-chain.
| Flag | Type | Required | Description |
|---|
--agent-uuid <uuid> | string | Yes¹ | Stored tokenized agent UUID |
--agent-id <id> | string | Yes¹ | On-chain ERC-8004 agent ID |
--metadata-key <key> | string | Yes² | Metadata key |
--metadata-value <value> | string | No | Metadata value |
--metadata-encoding <encoding> | string | No | string (default), json, or hex |
--ai-model-name <name> | string | No | Shortcut for metadataKey=modelName |
--ai-model-provider <provider> | string | No | AI model provider stored in local profile context |
--agent-uuid or --agent-id. ² Or supply --ai-model-name as the key shortcut.
brickken agent set-metadata \
--chain "$CHAIN" \
--signer-address "$WALLET" \
--agent-uuid "$AGENT_UUID" \
--metadata-key capabilities \
--metadata-value '{"tasks":["research","summarization","token-operations"]}' \
--metadata-encoding json \
--execute \
--json
For complex metadata payloads, prefer a --file to avoid shell-quoting issues. See the raw tx flow for the file-based pattern. agent set-wallet
Prepare or execute an agent wallet update → agentSetWallet. The new wallet must sign an authorization first.
| Flag | Type | Required | Description |
|---|
--agent-uuid <uuid> | string | Yes¹ | Stored tokenized agent UUID |
--agent-id <id> | string | Yes¹ | On-chain ERC-8004 agent ID |
--new-wallet <address> | string | Yes | New agent wallet address |
--signature <signature> | string | Yes | Authorization signature from the new wallet |
--deadline <value> | number | Yes | Unix timestamp deadline for the signature |
--agent-uuid or --agent-id.
brickken agent set-wallet \
--chain "$CHAIN" \
--signer-address "$WALLET" \
--agent-uuid "$AGENT_UUID" \
--new-wallet 0xNewAgentWallet \
--signature 0xSignatureFromNewWallet \
--deadline 1800000000 \
--execute \
--json
agent feedback give
Prepare or execute reputation feedback → agentGiveFeedback.
| Flag | Type | Required | Description |
|---|
--agent-uuid <uuid> | string | Yes¹ | Target agent UUID |
--agent-id <id> | string | Yes¹ | Target on-chain agent ID |
--value <value> | string | Yes | Feedback value |
--value-decimals <value> | string | Yes | Decimals used to interpret value |
--email <email> | string | No² | Feedback owner email |
--tag1 <value> | string | No | Feedback tag 1 |
--tag2 <value> | string | No | Feedback tag 2 |
--endpoint <url> | string | No | Service endpoint related to feedback |
--feedback-uri <uri> | string | No | Feedback URI |
--feedback-hash <hash> | string | No | 32-byte feedback hash (defaults to zero hash) |
--agent-uuid or --agent-id. ² Required when the API cannot infer tracking context from the agent reference.
brickken agent feedback give \
--chain "$CHAIN" \
--signer-address "$WALLET" \
--agent-id 7 \
--email reviewer@example.com \
--value 9500 \
--value-decimals 2 \
--tag1 quality \
--endpoint https://agent.example/.well-known/agent-card.json \
--execute \
--json
agent feedback revoke
Prepare or execute feedback revocation → agentRevokeFeedback.
| Flag | Type | Required | Description |
|---|
--agent-id <id> | string | Yes | On-chain ERC-8004 agent ID |
--feedback-index <index> | number | Yes | Feedback index to revoke |
--email <email> | string | Yes | Feedback owner email |
brickken agent feedback revoke \
--chain "$CHAIN" \
--signer-address "$WALLET" \
--agent-id 7 \
--feedback-index 0 \
--email reviewer@example.com \
--execute \
--json
agent feedback respond
Prepare or execute a feedback response → agentAppendFeedbackResponse.
| Flag | Type | Required | Description |
|---|
--agent-id <id> | string | Yes | On-chain ERC-8004 agent ID |
--client-address <address> | string | Yes | Original feedback client address |
--feedback-index <index> | number | Yes | Feedback index to respond to |
--response-uri <uri> | string | Yes | URI of the off-chain response payload |
--response-hash <hash> | string | No | 32-byte response hash (defaults to zero hash) |
--email <email> | string | Yes | Agent owner email |
brickken agent feedback respond \
--chain "$CHAIN" \
--signer-address "$WALLET" \
--agent-id 7 \
--client-address 0xReviewerWallet \
--feedback-index 0 \
--response-uri ipfs://QmResponseCID \
--email owner@example.com \
--execute \
--json
For multi-step agent flows (register → set-uri → set-metadata), wait for each transaction to be mined before preparing the next one. Preparing dependent transactions before the previous one is mined can produce nonce conflicts. See the end-to-end guide. 
