docs(v0): record D13 north-star + D14 OD7 dogfooding model (unblocks #7 KPI Tree) #39

Merged

pdurlej merged 1 commit from claude/patchwarden-decisions-d13-d14 into main

2026-05-26 11:16:10 +02:00

claude commented

2026-05-26 10:58:32 +02:00

Collaborator

Summary

Records two decisions that unblock issue #7 KPI Tree + OKRs (currently status:needs-operator-decision due to OD6 + OD7 pending). Both decisions are direct extracts from existing artifacts — D13 from kpi-tree.md §10 proposal, D14 from the implication of already-merged D6/D7. No new strategic claims; only formalizes what other docs already imply.

What changed

Summary table (docs/decisions.md § Summary table): adds rows D13 and D14 after D7. Numbering is non-sequential (D8–D12 reserved for future use; this matches strategy.md proposed D16 + kill-criteria.md proposed D17 numbering scheme — those will land in separate PRs).

New section D13 — North-star metric + Q3 OKR locking: selects (c) "PRs reviewed without false-blocker incidents — rolling 90-day count" (quality-led) as the v0 north-star, with a planned Q1 2027 switch to (a) "Paid Cloud users — monthly active" (revenue-led) at Cloud alpha. Rationale, risk-if-wrong, and what it settles all captured per existing decision template.

New section D14 — OD7 dogfooding default model: confirms that D6 (already merged 2026-05-25) settles OD7. Cloud-first for v0, local model deferred to post-v0, per-repo overrides via the cloud_review policy schema field (gap G9, to be implemented in Q3 per roadmap cousin-track).

Why these two specifically

Operator decision 2026-05-26: M1 close prep includes "PR z D13 north-star + D-future OD7 do decisions.md". These were the only two decisions explicitly approved for this batch. D16 (strategy lock) and D17 (kill-criteria cadence) remain proposed in their respective artifacts and will land in separate PRs to keep PR scope atomic.

What did NOT change

Existing D1–D7 untouched
Source attribution section untouched
No other docs touched
No new artifact files created

Downstream effect

Once merged, issue #7 flips from status:needs-operator-decision to status:cagan-grade-review-pending (both OD6 and OD7 now settled). Operator can then quality-bar review docs/product/kpi-tree.md and close #7 with the rest of the M1 quick-close batch.

Verification

grep -nE "^## D" docs/decisions.md shows D1–D7 + D13 + D14 (9 sections total)
Summary table has 9 rows
D13 specifies (c) as the v0 choice and (a) as Q1 2027 switch
D14 references D6 as the settling decision (not introducing a new model choice)
No edits outside docs/decisions.md

🤖 claude (Patchwarden dedicated thread), M1 close prep batch per operator decision 2026-05-26. Sibling deliverables: label cagan-grade-approved (id 121, created), follow-up issue #35 (created), PR #38 (open — product README M1 close prep), PR #34 (merged — roadmap restructure).

## Summary Records two decisions that unblock issue [#7 KPI Tree + OKRs](https://git.pdurlej.com/pdurlej/patchwarden/issues/7) (currently `status:needs-operator-decision` due to OD6 + OD7 pending). Both decisions are direct extracts from existing artifacts — D13 from `kpi-tree.md` §10 proposal, D14 from the implication of already-merged D6/D7. No new strategic claims; only formalizes what other docs already imply. ## What changed **Summary table** (`docs/decisions.md` § Summary table): adds rows D13 and D14 after D7. Numbering is non-sequential (D8–D12 reserved for future use; this matches `strategy.md` proposed D16 + `kill-criteria.md` proposed D17 numbering scheme — those will land in separate PRs). **New section D13 — North-star metric + Q3 OKR locking**: selects **(c) "PRs reviewed without false-blocker incidents — rolling 90-day count"** (quality-led) as the v0 north-star, with a planned Q1 2027 switch to (a) "Paid Cloud users — monthly active" (revenue-led) at Cloud alpha. Rationale, risk-if-wrong, and what it settles all captured per existing decision template. **New section D14 — OD7 dogfooding default model**: confirms that D6 (already merged 2026-05-25) settles OD7. Cloud-first for v0, local model deferred to post-v0, per-repo overrides via the `cloud_review` policy schema field (gap G9, to be implemented in Q3 per roadmap cousin-track). ## Why these two specifically Operator decision 2026-05-26: M1 close prep includes "PR z D13 north-star + D-future OD7 do decisions.md". These were the only two decisions explicitly approved for this batch. D16 (strategy lock) and D17 (kill-criteria cadence) remain proposed in their respective artifacts and will land in separate PRs to keep PR scope atomic. ## What did NOT change - Existing D1–D7 untouched - Source attribution section untouched - No other docs touched - No new artifact files created ## Downstream effect Once merged, issue [#7](https://git.pdurlej.com/pdurlej/patchwarden/issues/7) flips from `status:needs-operator-decision` to `status:cagan-grade-review-pending` (both OD6 and OD7 now settled). Operator can then quality-bar review `docs/product/kpi-tree.md` and close #7 with the rest of the M1 quick-close batch. ## Verification - [ ] `grep -nE "^## D" docs/decisions.md` shows D1–D7 + D13 + D14 (9 sections total) - [ ] Summary table has 9 rows - [ ] D13 specifies (c) as the v0 choice and (a) as Q1 2027 switch - [ ] D14 references D6 as the settling decision (not introducing a new model choice) - [ ] No edits outside `docs/decisions.md` --- 🤖 claude (Patchwarden dedicated thread), M1 close prep batch per operator decision 2026-05-26. Sibling deliverables: label `cagan-grade-approved` (id 121, created), follow-up issue #35 (created), PR #38 (open — product README M1 close prep), PR #34 (merged — roadmap restructure).

Rows
Columns