Brevet

io.github.BrightbeamAIv0.3.0Updated Oct 6, 2026

Change control for what AI agents learn: humans promote, releases are signed, rules can be recalled

VerifiedSTDIODesktop onlyAI & ML Developer Tools Security & Monitoring

Overview

AI-generated overview

Brevet wraps an existing AI agent to record human corrections, propose learned rules, require human approval, sign releases, and recall capabilities.

What it does: Brevet is a local-first Python runtime that wraps an agent you already have and governs its learning. It records expert corrections as overrides, turns recurring overrides into candidate rules during an offline dream cycle, and lets a named human or mission group promote them at a dawn gate. Promoted capabilities ship in a signed release listed in capabilities.lock, can be recalled later, and every step is recorded on a hash-linked evidence chain. It also replays overrides as evals and blocks a release that makes them worse.
When to use it: Use it when an agent changes its own memories, skills, or instructions and you need those changes to be written down, approved by an accountable person or group, and reversible. It suits teams that want an auditable record of what an agent learned and who approved it, without changing the agent itself.
Requirements: Runs locally as a Python package (brevet) via uvx or pip; no model or network connection is needed. Requires the BREVET_HOME environment variable pointing at a workspace folder holding agent.yaml and .brevet/, created on first use with brevet init. Optional BREVET_AUTO_CAPTURE controls automatic recording of corrections. Desktop only.

Before you install

It records corrections, approvals, releases, and recalls, so it writes governance data into the workspace. Promotions and releases are gated by named approvers, and a dream:* approver is rejected. Some protections, such as verifying who holds each key and anchoring the evidence chain outside the machine, depend on the deployment environment.

Installation

In SourceWeft

Open Brevet in the dashboard and add it to a workspace.
Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

[Brightbeam]

Brevet: Change Control for What AI Agents Learn

A governed evolution loop that makes agent learning promotable, auditable, and revocable.

[Python 3.10+] [License: Apache-2.0] [CI] [CHAP-compatible]

AI agents now change their own behaviour while they work. They save memories, write themselves new skills and edit their own instructions. Many of these changes help. Yet none of them passes through the steps an organisation expects when a person changes how work is done. Nobody writes the change down, nobody approves it, and when it turns out to be wrong there is no earlier version to go back to.

Brevet adds those steps. It is a local-first Python runtime that wraps the agent you already have and runs its learning as a governed evolution loop. When an expert corrects the agent's draft, Brevet records the correction and its reason as an override. Offline, in the dream cycle, overrides that keep recurring become candidate capabilities: proposed rules and other learned behaviour, with no authority. At the dawn gate, a named human or mission group (the accountable review board) decides which candidates to promote. The overrides then replay as evals, and the conservative gate stops a release that makes either half of them worse. Promoted capabilities ship in a signed release, listed in capabilities.lock, and a capability that proves wrong can be recalled, with every release that shipped it flagged. Every step is recorded on a hash-linked evidence chain.

Agents propose deltas; evidence tests them; humans promote them; the runtime only ever executes signed versions.

[The governed evolution loop: the agent works under one signed harness; an expert's correction is recorded as an override; the dream cycle turns recurring overrides into candidate capabilities with no authority; at the dawn gate a named human or mission group promotes, holds or rejects each candidate; evals replay the overrides and the conservative gate must pass; promoted capabilities ship in a signed release listed in capabilities.lock. A capability that proves wrong is recalled and every release that shipped it is flagged.]

Three questions Brevet answers

Question	How Brevet answers it
What has the agent learned?	Every release carries `capabilities.lock`, the capability bill of materials: each learned capability, where it came from and the hash of its exact content.
Who approved it?	Each capability records the human or mission group that promoted it at the dawn gate, with the overrides that justified it.
How do we take it back?	Recall it. Brevet flags every release that shipped it and records why it was recalled.

A concrete example

A quality reviewer at a pharmaceutical plant checks an agent's severity rating for each equipment problem. The agent rates pump vibration during cleaning as minor. She overrides it to major every time, because that vibration is an early sign of seal wear. After four overrides, the dream cycle proposes a candidate rule. At dawn her mission group promotes it, and release 0.2.0 ships with the rule in its capabilities.lock. Months later, engineers trace the vibration to a faulty sensor, so the mission group recalls the rule and Brevet flags release 0.2.0.

[The example as a timeline: over three weeks the reviewer records four overrides; that night the dream cycle proposes a candidate rule at the Evidence layer; next morning at dawn the mission group promotes it and release 0.2.0 ships signed; months later the rule is recalled and release 0.2.0 is flagged.]

The same story in code

The agent here is a plain Python function; with a real framework you pass your agent object instead.

python

import brevetfrom brevet.runner import EvalRunner
agent = brevet.wrap(triage_agent)          # wrap the agent you already have
# Work and override: the agent drafts; the reviewer corrects the draft and says why.result = agent.run("Pump P-301: vibration high during cleaning",                   task_family="equipment_triage")agent.record_final(result.task_id, "severity: major",                   participant="human:[email protected]",                   rationale="Vibration during cleaning is an early sign of seal wear.",                   tags=["vibration-during-cleaning"])
# Dream: once the same override keeps recurring, it becomes a candidate.agent.dream()candidates = agent.dawn()                   # the dawn queuerule = next(c for c in candidates if c.kind == "prompt_rule")
# Dawn: a named mission group promotes it. A dream:* approver is rejected.agent.dawn(decide=(rule.capability_id, "promote"),           approver="mission_group:quality_team")
# Evals: replay the overrides as tests, before and after the change.before = agent.evaluate()#    ...update your agent so that it follows the promoted rule...after = agent.evaluate()
# Release: refused unless the conservative gate passes.check = EvalRunner.compare(before, after)agent.release(to_version="0.2.0", channel="trial",              approver="mission_group:quality_team",              delta_in=check["delta_held_in"], delta_out=check["delta_held_out"])
# Recall: the rule proves wrong. Then verify the whole evidence chain.agent.recall(rule.capability_id, reason="The vibration came from a faulty sensor.",             issued_by="mission_group:quality_team")agent.verify()

import brevetfrom brevet.runner import EvalRunner
agent = brevet.wrap(triage_agent)          # wrap the agent you already have
# Work and override: the agent drafts; the reviewer corrects the draft and says why.result = agent.run("Pump P-301: vibration high during cleaning",                   task_family="equipment_triage")agent.record_final(result.task_id, "severity: major",                   participant="human:[email protected]",                   rationale="Vibration during cleaning is an early sign of seal wear.",                   tags=["vibration-during-cleaning"])
# Dream: once the same override keeps recurring, it becomes a candidate.agent.dream()candidates = agent.dawn()                   # the dawn queuerule = next(c for c in candidates if c.kind == "prompt_rule")
# Dawn: a named mission group promotes it. A dream:* approver is rejected.agent.dawn(decide=(rule.capability_id, "promote"),           approver="mission_group:quality_team")
# Evals: replay the overrides as tests, before and after the change.before = agent.evaluate()#    ...update your agent so that it follows the promoted rule...after = agent.evaluate()
# Release: refused unless the conservative gate passes.check = EvalRunner.compare(before, after)agent.release(to_version="0.2.0", channel="trial",              approver="mission_group:quality_team",              delta_in=check["delta_held_in"], delta_out=check["delta_held_out"])
# Recall: the rule proves wrong. Then verify the whole evidence chain.agent.recall(rule.capability_id, reason="The vibration came from a faulty sensor.",             issued_by="mission_group:quality_team")agent.verify()

The full script is examples/pump_vibration.py, and ABOUT.md shows what it prints.

Quickstart

console

pip install brevetbrevet demo                  # the whole loop as one commandbrevet playground            # step through the loop in your browser

pip install brevetbrevet demo                  # the whole loop as one commandbrevet playground            # step through the loop in your browser

Everything runs on your own machine, with no model or network connection. To run the example above, clone the repository and run python examples/pump_vibration.py. For a guided, clickable tour, open docs/demo.html in a browser.

Works with the agent you already have

brevet.wrap() recognises agents built with LangGraph, the Claude Agent SDK, DeepAgents, AutoGen, LlamaIndex, Pydantic AI, the Google Agent Development Kit, CrewAI and the OpenAI Agents SDK, and it accepts any Python function. Brevet never changes the agent it wraps. uvx brevet mcp offers the whole loop to any MCP client (ABOUT.md shows the setup), and examples/claude-cowork uses it to govern what Claude itself learns.

Project status

Brevet implements the whole loop and keeps every record, and a workspace can require every decision to be signed by its registered approvers. Some protections depend on the system you deploy it in, such as verifying who holds each key and anchoring the evidence chain outside the machine. ABOUT.md lists them, and the paper sets them out in full.

Learn more

ABOUT.md: the seven stages, the authority ladder, supported frameworks, the MCP server and commands, how Brevet fits with CHAP and Metis, and how the repository is organised.
GLOSSARY.md: every term, with its plain meaning first.
SPEC.md: the rules any implementation must follow.
BENCHMARK.md: the proposed governed-adaptation benchmark.

Citation

If you use Brevet in research, please cite the paper Brevet: Change Control for What Self-Evolving AI Agents Learn (Shahid, Suttie and Black, 2026). CITATION.cff gives the software citation.

License

Apache-2.0. See LICENSE. Brevet is a Brightbeam project.

Source: README.md at commit c4b8a19

Tools

Tool metadata has not been indexed yet.

Version history

v0.3.0LatestOct 6, 2026