McRitchie Studio
Say Hi 👋
Agents Builders
← All Tasks
Sizing

Conductor refetch queue rule

Archived
conductor-refetch-queue-rule

Created

Jun 25, 05:49

Started

Jun 25, 05:50

Completed

Jun 25, 12:23

DevOps handoff

Type

Docs

Shape

—

Worktree Slug

conductor-refetch-queue-rule

Repositories

mcritchie-studio

Release Train

—

Branch

feat/conductor-refetch-queue-rule

Pull Request

https://github.com/amcritchie/mcritchie-studio/pull/187

Local URL

—

QA URL

—

Production URL

—

docs agent-ops

Acceptance Criteria

  • SOP says refetch submitted queue before each act
  • covers before every merge and every deploy

Expected Test Plan

No expected checks recorded.

Checks Run

  • [review] Shannon (heavy/ui) + Alex (light/docs) APPROVE_WITH_NITS — both acceptance criteria met in parallel-agent-devops.md; all technical claims (merge reopens candidate, ship deploys frozen qa_shas + runs prod hooks, merge-tree conflict check, mergeable=UNKNOWN after base change) verified accurate. FOLLOWUP: builder claimed to also edit canonical devops-cycle-design.md (Full spec) but did not — add the standing rule or a cross-reference there.

Agent Context

From conductor retro 2026-06-25. Doc-only (kind=docs, dor-check exempt). Add to the conductor SOP (docs/agents/system/devops-cycle-design.md + docs/agents/modules/parallel-agent-devops.md) the standing rule: the board is continuously fed by builders + other sessions, so the conductor MUST re-fetch 'bin/task list --stage submitted' immediately before every merge and every deploy — never act on a snapshot older than your own last write (this session shipped a 1-PR release on a stale 3-PR-stale snapshot). Also note: after merging into release, other open PRs flip to GitHub mergeable=UNKNOWN (async recompute); do NOT trust gh mergeable right after a base change — use 'git merge-tree --write-tree origin/release origin/<branch>' as the reliable local conflict check.

Stage Timeline

Who handled each stage, the time it took (measured), and the model / tokens / cost reported (best-effort) — plus who's on it right now. — means the agent didn't report that metric.

  1. Created → Designed
    N Ninetales
    Ninetales
    Model
    —
    Duration
    —
    Tokens
    —
    Cost
    —
    Completed Jun 25, 05:49 · 2 days ago
    api
  2. Designed → Building
    N Ninetales
    Ninetales
    Model
    claude-opus-4-8
    Duration
    2 minutes
    Tokens
    —
    Cost
    —
    Started Jun 25, 05:49
    Completed Jun 25, 05:50 · 2 days ago
    cli
  3. Building → Submitted
    N Ninetales
    Ninetales
    Model
    claude-opus-4-8
    Duration
    2 minutes
    Tokens
    8,811,171
    Cost
    ~$5.05
    Started Jun 25, 05:50
    Completed Jun 25, 05:52 · 2 days ago
    cli
  4. Submitted → Reviewed
    S Shannon
    Shannon primary
    A Alex
    Alex light
    Model
    claude-opus-4-8
    Duration
    8 minutes
    Tokens
    —
    Cost
    —
    Started Jun 25, 05:52
    Completed Jun 25, 06:01 · 2 days ago
    cli
  5. Reviewed → Assembled
    S Steffon
    Steffon
    Model
    —
    Duration
    7 minutes
    Tokens
    —
    Cost
    —
    Started Jun 25, 06:01
    Completed Jun 25, 06:07 · 2 days ago
  6. Assembled → Shipped
    A Avi
    Avi
    Model
    —
    Duration
    about 6 hours
    Tokens
    —
    Cost
    —
    Started Jun 25, 06:07
    Completed Jun 25, 12:23 · 2 days ago
  7. Shipped → Archived
    8
    81fad33e-98b5-4075-8ed6-d0efe6cb5804
    Model
    claude-opus-4-8
    Duration
    about 9 hours
    Tokens
    —
    Cost
    —
    Started Jun 25, 12:23
    Completed Jun 25, 21:16 · 1 day ago
    cli

Conversation

QA review feedback, agent handoffs, and follow-up notes for this task.

No task feedback has been recorded yet.

Sealed-bid sizing

Edit →

Alex (PM)

—

Avi (PO)

—

Dev

—

Actual

—