neutral

Declarative reconcilers that take out-of-draft LLPs and converge a repo all the way to review-ready PRs — technical design → implementation design → task fan-out → integration PR → review → fix.

Each stage is a reconciler, not a script. It holds a base state (an invariant) and closes the gap from git/file ground truth. It can't fabricate "done", because "done" is a fact you re-derive from the world — a real commit on the integration branch, a real @ref in a design — never a status field an agent wrote.

out-of-draft LLPs ──Designer──▶ ordered change sets (a DAG), each with a
                                technical design that @refs the LLPs it covers
        change set ──ImplDesigner──▶ implementation design + task list (dep edges)
        change set ──Implementer──▶ wave loop: parallel impl (worktree-isolated)
                                    then SERIAL verified merge into integration/<cs>
        change set ──Reviewer──▶ /dual-review → fix loop (≤N) → HOLD ready-for-merge

Why declarative

The predecessor (/feature-flow) was imperative formulas. It worked, but its failure modes were all the same shape: agents fabricating results under ambiguous prose — a merge recorded with a non-existent SHA, a review verdict invented, a recovery step that threw away completed work. Neutral's reconcilers read ground truth before acting and verify every claim against it, so those failures can't survive. See LLP 0001 and LLP 0002.

The ready-queue, without beads

The one indispensable thing a task tracker gives you is "list the unblocked open tasks." Neutral derives it from git instead of a database:

• dependency edges are declared in the change set's implementation plan LLP (where the impl-designer already decides ordering),
• completion is derived from git: a task is done iff its branch is a verified ancestor of the integration branch.

ready = { t : not done(t) and deps(t) ⊆ done } — a pure function of (task file, git state). No ledger to drift or fabricate. See LLP 0003.

Status

Optional PR backlog audit

Enable the advisory autophagy member in each repo's .neutral/config.json, merging this field into its existing configuration:

{ "autophagy": { "prBacklog": true } }

On an eligible idle tick, the loop assesses all open PRs for remaining need, already-landed work, supersession, or changed requirements. Evidence and proposed human actions go into a held draft report PR; assessed PRs are left unchanged. An unmerged replacement is a dependency, and uncertain cases remain open. The option defaults off and shares autophagy's admission gate, cooldowns and one-open-proposal limit, so busy repos may defer the audit.

neutral idle --json exposes the selection, eligibility reasons and complete PR inventory. Complete no-op scans use --backlog-snapshot <fingerprint> to avoid repeats until the observed backlog changes. See LLP 0079 for the evidence and delivery rules.

Deploy the updated CLI and reconcile skill before enabling the option in the production repos. Verify each repo's neutral idle --json includes a pr-backlog member and a successful prBacklog observation; initiative: null is expected when existing work, admission, cooldowns or an open autophagy PR blocks selection.

Development

Milestone 0: deterministic core + neutral status. The reconcilers (Designer, Impl-designer, Implementer, Reviewer) land in M1–M4.

neutral status   # corpus by stage, coverage gap, change-set readiness
npm test