Agents Builders

Action grade and insights

Archived
action-grade-and-insights

Created

Jun 30, 20:19

Started

Jun 30, 20:23

Completed

Jun 30, 23:22

DevOps handoff

Type

Feature

Shape

backend

Worktree Slug

action-grade-and-insights

Repositories

mcritchie-studio

Release Slug

Branch

feat/action-grade-and-insights

Local URL

QA URL

Production URL

migration

Acceptance Criteria

  • Store Alex grade slug disposition longform
  • Store McRitchie audit of Alex
  • Bank or discard each graded action
  • Insight bank holds banked grades
  • Disposition is binary good or not

Expected Test Plan

  • [unit] grade model validations and disposition
  • [integration] banking and discarding a grade

Checks Run

  • [unit] grader/disposition inclusion enums; slug presence; atomic_action required
  • [unit] uniqueness one grade per (atomic_action, grader); same grader on other action ok
  • [unit] bank!/discard! mutual exclusivity; banked/discarded defaults false; predicates
  • [unit] scopes by_grader/banked/discarded/for_action
  • [integration] bank a grade -> appears in Insight Bank scope; discard -> excluded
  • [integration] alex grade + mcr audit two rows on one action; dependent: destroy cascade
  • [full-suite@9227ae1a8d604b1fe4bb9a5e5393c7f6edb76eb4] bin/rails test green (clean-CI parity, CLAUDE_CODE_SESSION_ID unset); live worktree 3 failures = known agent-marker env issue, not this change
  • [rubocop@9227ae1a8d604b1fe4bb9a5e5393c7f6edb76eb4] bin/rubocop clean (591 files, no offenses)

Agent Context

Grading layer over AtomicAction (in release). Model Alex's grade + the McRitchie audit-of-Alex: 4-7 word feedback slug, binary disposition (good/not), optional long-form, plus banked/discarded flags. Recommend an ActionGrade model belongs_to AtomicAction with grader role (alex|mcr) so the two layers are two rows; Insight Bank = banked grades. Reference scratchpad atomic-heartbeat.html (the a/m feedback objects: sl, d, lg, banked, discarded). No UI; the heartbeat UI wiring is a separate follow-up (T4).

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.

Sizing Avi · PO MEDIUM Dev SMALL Actual XL ≠ forecast
  1. Created Designed
    C Caterpie
    Caterpie
    Model
    Duration
    Tokens
    Cost
    Completed Jun 30, 20:19 · 8 days ago
    api
  2. Designed Building
    C Caterpie
    Caterpie
    Model
    claude-opus-4-8
    Duration
    4 minutes
    Tokens
    11,847,181
    Cost
    ~$10.79
    Started Jun 30, 20:19
    Completed Jun 30, 20:23 · 8 days ago
    cli
  3. Building Submitted
    C Caterpie
    Caterpie
    Model
    claude-opus-4-8
    Duration
    15 minutes
    Tokens
    20,069,309
    Cost
    ~$18.58
    Started Jun 30, 20:23
    Completed Jun 30, 20:38 · 8 days ago
    cli
  4. Submitted Reviewed Review Activity
    C Carl
    Carl primary
    A Alex
    Alex light
    Model
    claude-opus-4-8
    Duration
    under a minute
    Tokens
    Cost
    Started Jun 30, 20:38
    Completed Jun 30, 20:39 · 8 days ago
    cli
  5. Reviewed Assembled
    S Steffon
    Steffon
    Model
    claude-opus-4-8
    Duration
    11 minutes
    Tokens
    6,807,913
    Cost
    ~$4.41
    Started Jun 30, 20:39
    Completed Jun 30, 20:49 · 8 days ago
  6. Assembled Shipped
    A Avi
    Avi
    Model
    claude-opus-4-8
    Duration
    about 3 hours
    Tokens
    68,712,028
    Cost
    ~$44.53
    Started Jun 30, 20:49
    Completed Jun 30, 23:22 · 8 days ago
  7. Shipped Archived
    Model
    Duration
    3 days
    Tokens
    Cost
    Started Jun 30, 23:22
    Completed Jul 04, 03:23 · 4 days ago

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)

MEDIUM

Dev

SMALL

Actual

XL