Sweeps

Mon, 01 Jan 0001 00:00:00 +0000

Use this page when the question is about the active sweep, the current anchor, or the next allowed research change.

A sweep is the repo’s way of testing one research change at a time without losing attribution. If you need the broader architecture or synthetic-data context, pair this page with Research.

Start With

Sweep Contract: queue discipline, locked surface, artifact requirements, and execution loop
Workflows: command syntax for inspect, execute, promote, render, and validate
Roadmap: which questions are actually next
References: evidence and literature context

Inspect-First Commands

.venv/bin/tab-foundry research sweep list-sweeps
.venv/bin/tab-foundry research sweep list --sweep-id <sweep_id>
.venv/bin/tab-foundry research sweep next --sweep-id <sweep_id>
.venv/bin/tab-foundry research sweep summarize --sweep-id <sweep_id> --include-screened
.venv/bin/tab-foundry research sweep inspect --order <order> --sweep-id <sweep_id>

Agent Program

Mon, 01 Jan 0001 00:00:00 +0000

Use this contract when you are running or reviewing a selected anchor-only system-delta sweep in tab-foundry.

Use docs/workflows.md for command syntax and artifact expectations. Use this contract for the objective, the locked comparison surface, the queue discipline, and the interpretation policy.

Treat this as a research-execution contract, not the architecture roadmap. A selected sweep may intentionally hold a PFN-adjacent, hybrid diagnostic, or sandwich-local surface fixed while isolating one question. The long-term direction for the public model surface still comes from docs/development/roadmap.md and docs/development/model-architecture.md.

Research Contributors on tab-foundry Docs

Sweeps

Start With

Inspect-First Commands

Agent Program