Local-first OSS for human-agent work

Governed continuity for human-agent work.

Useful context should survive. Authority should be ratified. Govinuity turns agent work into review-gated, measurable continuity: surface candidate decisions, ratify what governs, inject only eligible context, and observe what happened after reuse.

Surface
candidate signal
Ratify
human authority
Measure
measured reuse

The protocol

Action, artifact, consequence.

Govinuity is not a place to store everything. Each step produces a visible artifact so continuity can be reviewed, reused, and measured.

Surface

Surface candidate decisions from work.

Harvest reads pasted transcripts, handoff summaries, local session files, or direct API submissions. It looks for decisions and continuity signals that may deserve to survive.

The output is not memory. It is a reviewable proposal with a reason for surfacing.

Harvest receipt
$ harvest --input session.txt --source codex
3 candidates surfacedreview queue
1 correction signal loggedruns
2 duplicates ignoreddedupe

Candidate decisions move forward with rationale, scope, reversibility, confidence, and provenance.

Review

Review the proposal before it becomes authority.

The review card is the contract surface. It shows what was proposed, why it was surfaced, how reversible it is, and where it should apply.

This is the governance gate: candidates stay non-authoritative until a human acts.

Review card

All DB migrations must be human-reviewed before production.

Why surfaced: explicit correction after a failed migration path
Class: durable constraint
Scope: project · Confidence: 88%
Reversibility: low — bad migrations are hard to unwind

Ratify

Ratify what becomes reusable context.

Approve, defer, reject, or supersede. Only approved decisions become eligible for future injection. Everything else remains outside the governing context.

Ratification is not just a button; it changes the authority state of the proposal.

Ratification receipt
Statusapproved
Scopeproject
Transfer tierby_project
Injection eligibilityeligible

Ratified decisions can govern future sessions. Deferred, rejected, and superseded proposals cannot.

Inject

Inject only eligible continuity.

Agents can request scoped continuity through GET /api/memory or read a generated GOVERNED_CONTINUITY.md file.

Every injection also records what was excluded. Stale, superseded, expired, or out-of-scope decisions do not silently leak forward.

Memory injection summary
GET /api/memory?project=api
2 decisions injectedactive
1 excludedproject mismatch
1 excludedsuperseded

The returned context is scoped, ratified, and logged as a continuity run.

Measure

Measure what happened after reuse.

Continuity runs make reuse inspectable: what was injected, what was excluded, and which outcome signals appeared later.

Signals are evidence for human interpretation, not fake certainty. They show where continuity helped, drifted, or needs revision.

Run outcome record
run-1744…k3f9a
project-api · 2 injected · 0 excluded
decision followed
run-1744…m7c2x
project-web · 1 injected · 1 excluded
correction required

Product console

A local governance console — no cloud, no accounts.

The product is intentionally operational: proposal intake from work, a review queue for incoming candidates, a decision log for ratified context, and run history for measuring reuse.

Govinuity
Dashboard Harvest Review Decisions Runs
Runs Injection history — what was injected, what was excluded, and outcome signals.
Continuity runs become the audit trail.
Injected decisions 0
Excluded as stale or scoped out 0
git clone https://github.com/Taprooted/Govinuity.git
cd Govinuity
npm install
cp .env.example .env.local
npm run dev

Quickstart

Run it locally.

Start the Next.js dashboard, submit a proposed decision, ratify it, then generate or request governed context for future agent sessions.

Current public release

Early, local-first, and built for technical users.

  • Local-first software with no accounts and no telemetry.
  • The full pipeline is in the public repo: proposal intake, review, ratified decisions, memory injection, run logging, and annotations.
  • Technically literate users can run it themselves, inspect the SQLite-backed data, and integrate it with agent workflows.
  • Proposal surfacing is core to the system. The Harvest page lets you scan session files, paste conversation text from any agent tool, or run the harvest script directly — including automatic correction signal detection and run annotation.