Logos Autonomous Agent: Difference between revisions

From Local Area Network
New tool page, source-checked against the project README
 
Plainer prose: fewer quotes, project claims attributed, glosses folded in
 
(One intermediate revision by the same user not shown)
Line 10: Line 10:
| verified_against = README and commit of 9 September 2026
| verified_against = README and commit of 9 September 2026
}}
}}
'''Logos Autonomous Agent''' is a library and headless program that gives an AI agent its own wallet and tools, and it runs on the Logos Execution Zone (LEZ), [[Logos Storage]] and [[Logos Messaging]], with a Basecamp owner app. The code is published by the GitHub account aegonmyy. The README does not say whether it is an official Logos project.
'''Logos Autonomous Agent''' is a library and headless program by the GitHub account aegonmyy that gives an AI agent its own wallet and tools. It runs on the Logos Execution Zone (LEZ), the Logos layer that runs on-chain programs, and on [[Logos Storage]] and [[Logos Messaging]], with a Basecamp owner app. The README does not say whether it is an official Logos project.


== What it does ==
== What it does ==
* The agent holds its own shielded LEZ account and can send and receive tokens apart from its owner.
* The agent holds its own shielded LEZ account and can send and receive tokens apart from its owner.
* The owner sets a per-transaction limit and a per-period limit. Below them the agent spends on its own. Above them it asks the owner over an encrypted channel and waits.
* The owner sets a per-transaction limit and a per-period limit. Within both limits the agent spends on its own, and over either limit it asks the owner over an encrypted channel and waits.
* Only the owner can change the spending limits. The agent cannot change its own.
* The README says only the owner can change the spending limits, and that the agent cannot change its own.
* Skills are called by name with JSON arguments. New skills, including third-party ones, register without changes to the agent core.
* Skills are called by name with JSON arguments, and new ones, including third-party skills, register without changes to the agent core.
* Default skills cover storage, messaging, the wallet and programs, A2A tasks and status.
* Default skills cover storage, messaging, the wallet and programs, status, and A2A tasks. A2A (Agent2Agent) is an open protocol for AI agents to find and work with each other.
* Agents publish signed A2A Agent Cards, find each other on a shared topic, and pay each other in LEZ per task. Logos Messaging carries the traffic.
* Agents publish signed A2A Agent Cards, find each other on a shared topic and pay each other in LEZ per task, with Logos Messaging carrying the traffic.
* Files are encrypted on the client before upload, so the storage node holds only ciphertext.
* By the README's account, files are encrypted on the client before upload, so the storage node holds only ciphertext.
* No AI model is bundled. The README says inference is left to the deployer.
* No AI model is bundled, and inference is left to the deployer.


Stage: Prototype. The README states no stage, and it points to a recorded run of the whole sequence against the public testnet. It lists known limitations, including that the wallet-backed on-chain skills are not yet driven through the Logos Core module session.
Stage: Prototype. The README states no stage. It points to a recorded run of the whole sequence against the public testnet, which <code>docs/HEADLESS_DEPLOY.md</code> dates 8 September 2026. The metadata gives version 1.0.0, and the latest tag is v0.2.0-basecamp-bundles. The README lists known limitations, including that the wallet-backed on-chain skills are not yet driven through the Logos Core module session.


The README says it was "Built for λ-Prize LP-0008".
The README says it was built for λPrize LP-0008.


== Use it ==
== Use it ==
# Set <code>LEE_WALLET_HOME_DIR</code> to a wallet folder. On a machine with no wallet, create one with <code>agent-wallet-init --sequencer https://testnet.lez.logos.co</code>.
# Set <code>LEE_WALLET_HOME_DIR</code> to a wallet folder. On a machine with no wallet, create one with <code>agent-wallet-init --sequencer https://testnet.lez.logos.co</code>.
# Start the agent with the <code>agent</code> command. Give it the owner identity, the spending limits, the nwaku REST address for Logos Messaging and a state file.
# Start the agent with the <code>agent</code> command. Give it the owner identity, the spending limits, a state file and the REST address of nwaku, a Logos Messaging node.
# The agent prints its shielded account ID. Fund it from any wallet.
# The agent prints its shielded account ID. Fund it from any wallet.
# The owner approves or denies held spends and changes limits over the owner channel, either in code with <code>OwnerChannel</code> or in the Basecamp owner app.
# The owner approves or denies held spends and changes limits over the owner channel, either in code with <code>OwnerChannel</code> or in the Basecamp owner app.
Line 34: Line 34:


== Build it ==
== Build it ==
* Run <code>nix build .#owner-install</code> for the owner app module and <code>nix build .#agent-install</code> for the core agent module.
* <code>nix build .#owner-install</code> builds the owner app module, and <code>nix build .#agent-install</code> builds the core agent module.
* Run <code>./scripts/build-ffi.sh</code> to build the Rust core library.
* <code>./scripts/build-ffi.sh</code> builds the Rust core library.
* Run <code>./scripts/package-basecamp.sh</code> to make side-loadable .lgx bundles in dist/.
* <code>./scripts/package-basecamp.sh</code> makes side-loadable .lgx bundles in dist/.
* Run the tests with <code>cargo test -p logos_agent</code>. The integration tests start a LEZ stack through Docker.
* <code>cargo test -p logos_agent</code> runs the tests. The integration tests start a LEZ stack through Docker.


== Releases ==
== Releases ==
* Latest tag: v0.2.0-basecamp-bundles, 9 September 2026.
* Latest tag: v0.2.0-basecamp-bundles, 9 September 2026.
* Version in metadata.json: 1.0.0, for both the core module (agent) and the owner app (agent_owner).
* Version in metadata.json: 1.0.0, for both the core module (agent) and the owner app (agent_owner).
* GitHub lists two releases. The latest is v0.2.0-basecamp-bundles, which carries agent.lgx, agent_owner.lgx and liblogos_agent.so.
* GitHub lists two releases. The latest is v0.2.0-basecamp-bundles, which carries agent.lgx, agent_owner.lgx, liblogos_agent.so, README.txt and SHA256SUMS.txt.


== See also ==
== See also ==
Line 56: Line 56:
* [https://github.com/aegonmyy/logos-agent/blob/main/module/metadata.json module/metadata.json], GitHub
* [https://github.com/aegonmyy/logos-agent/blob/main/module/metadata.json module/metadata.json], GitHub
* [https://github.com/aegonmyy/logos-agent/blob/main/app/metadata.json app/metadata.json], GitHub
* [https://github.com/aegonmyy/logos-agent/blob/main/app/metadata.json app/metadata.json], GitHub
* [https://github.com/aegonmyy/logos-agent/blob/main/docs/HEADLESS_DEPLOY.md docs/HEADLESS_DEPLOY.md], GitHub
* [https://github.com/aegonmyy/logos-agent/releases releases], GitHub


{{Page status|Stub}}
{{Page status|Stub}}

Latest revision as of 23:27, 4 November 2026

Logos Autonomous Agent
What it does: A module that gives an AI agent its own shielded LEZ wallet, private file storage, encrypted messaging, an owner-controlled spending policy and A2A-compatible coordination
Stage: Prototype
Licence: MIT or Apache-2.0
Verified on 4 November 2026 against README and commit of 9 September 2026.

Logos Autonomous Agent is a library and headless program by the GitHub account aegonmyy that gives an AI agent its own wallet and tools. It runs on the Logos Execution Zone (LEZ), the Logos layer that runs on-chain programs, and on Logos Storage and Logos Messaging, with a Basecamp owner app. The README does not say whether it is an official Logos project.

What it does

  • The agent holds its own shielded LEZ account and can send and receive tokens apart from its owner.
  • The owner sets a per-transaction limit and a per-period limit. Within both limits the agent spends on its own, and over either limit it asks the owner over an encrypted channel and waits.
  • The README says only the owner can change the spending limits, and that the agent cannot change its own.
  • Skills are called by name with JSON arguments, and new ones, including third-party skills, register without changes to the agent core.
  • Default skills cover storage, messaging, the wallet and programs, status, and A2A tasks. A2A (Agent2Agent) is an open protocol for AI agents to find and work with each other.
  • Agents publish signed A2A Agent Cards, find each other on a shared topic and pay each other in LEZ per task, with Logos Messaging carrying the traffic.
  • By the README's account, files are encrypted on the client before upload, so the storage node holds only ciphertext.
  • No AI model is bundled, and inference is left to the deployer.

Stage: Prototype. The README states no stage. It points to a recorded run of the whole sequence against the public testnet, which docs/HEADLESS_DEPLOY.md dates 8 September 2026. The metadata gives version 1.0.0, and the latest tag is v0.2.0-basecamp-bundles. The README lists known limitations, including that the wallet-backed on-chain skills are not yet driven through the Logos Core module session.

The README says it was built for λPrize LP-0008.

Use it

  1. Set LEE_WALLET_HOME_DIR to a wallet folder. On a machine with no wallet, create one with agent-wallet-init --sequencer https://testnet.lez.logos.co.
  2. Start the agent with the agent command. Give it the owner identity, the spending limits, a state file and the REST address of nwaku, a Logos Messaging node.
  3. The agent prints its shielded account ID. Fund it from any wallet.
  4. The owner approves or denies held spends and changes limits over the owner channel, either in code with OwnerChannel or in the Basecamp owner app.
  5. For the Basecamp app, set the environment variables the README lists, then load the agent and agent_owner modules into logoscore or Basecamp.

Build it

  • nix build .#owner-install builds the owner app module, and nix build .#agent-install builds the core agent module.
  • ./scripts/build-ffi.sh builds the Rust core library.
  • ./scripts/package-basecamp.sh makes side-loadable .lgx bundles in dist/.
  • cargo test -p logos_agent runs the tests. The integration tests start a LEZ stack through Docker.

Releases

  • Latest tag: v0.2.0-basecamp-bundles, 9 September 2026.
  • Version in metadata.json: 1.0.0, for both the core module (agent) and the owner app (agent_owner).
  • GitHub lists two releases. The latest is v0.2.0-basecamp-bundles, which carries agent.lgx, agent_owner.lgx, liblogos_agent.so, README.txt and SHA256SUMS.txt.

See also

Sources

Page status: Stub. Stub · Outline · Usable · Guide · Star. Anyone can improve this page and change its status.
Retrieved from "https://alienware-1.tailb5c716.ts.net/index.php?title=Logos_Autonomous_Agent&oldid=856"