Skip to content

TEMPLATE.md

Latest commit

 

History

History
88 lines (62 loc) · 4.06 KB

TEMPLATE.md

File metadata and controls

88 lines (62 loc) · 4.06 KB
Raw

Cycle NNNN — Outcome

Type: experiment | product-increment (feature cycle) | bug-fix cycle | hardening
Status: ready
Result: pending
Delivery: not-applicable
Goal version: copy the current PRODUCT_GOAL.md version
Tracking issue: link after cycle selection
Goal issue: https://github.com/laststance/codebook/issues/1\ Goal links: Gxx
Source state: link to preceding cycle or initial PROJECT_STATE.md

Question and authorized scope

One primary question. Why does resolving it advance the product now?

Hypothesis and scope

  • Hypothesis:
  • Smallest observable outcome:
  • In scope:
  • Out of scope for this cycle (still retained in PRODUCT_GOAL.md):
  • Conditions that falsify the approach:

Plan and selected skills

Draft only this cycle's executable steps. List the GStack skills needed and why; do not require every skill by default. Link their actual reports and record reviewed plan/version. Record authority for any external delivery operations.

Accepted product and design

  • Authorized phase and selected acceptance-map IDs:
  • Approved PNG path / hash / approval status from DESIGN.md:
  • Image-viewing action actually performed and observed live/replay/Wiki/shortcut composition:
  • Relevant difference between the current implementation and that visual target:
  • Linked full feature contracts / controlling owner decisions:

Open the actual PNG; metadata or a prompt is not inspection. Use the evidence template for the detailed receipt and each host's observations.

Source-gap audit and selected increment

  • Baseline source/goal/image identity; all-row or affected-row audit and why:
  • Every AC accounted for in PROJECT_STATE; inspected entry/symbol and implementation mismatch / UNKNOWN separately from proof:
  • Primary gap and dependency/regression AC IDs; why selected now:
  • Source change or bounded experiment; starting action → expected outcome; falsifying observation:
  • Carry-forward rationale for unaffected rows:

A read receipt or file's existence cannot satisfy this audit. UNKNOWN requires a concrete inspection action. Update affected rows after implementation/verification.

Carried defects

A bug-fix cycle only (a feature cycle starts with none and removes this section): every defect the previous cycle left unfixed, copied in whole from its record under the cycle rule, and every defect found while this cycle runs. None is optional and none is carried again; the cycle closes only when every row is closed.

Source (cycle, finding label) Fix, or the evidence that closes it Test that fails without the fix Closed

Validation contract

Freeze expectations before evaluating the implementation. Fill versions, actual targets, exact real-project/source binding (identify isolated mechanism-test inputs separately), setup, and commands while planning.

Check Expected observable result Real procedure Evidence required

Attempts and evidence

No observations yet. For each attempt: date, source revision/dirty state, versions, action, expected result, actual result, artifact/availability, limits. Label mock, mechanism-test, real-repository and packaged evidence accurately; bundled target demos cannot establish practical acceptance.

Result and decision

  • Result: pending / demonstrated / falsified / inconclusive
  • Findings:
  • Keep / revise / replace:
  • Evidence that can update goal state:
  • Acceptance-map IDs advanced / actual UI compared with the approved image:
  • Before/after source and proof gaps for every affected AC; regressions and unaffected-evidence rationale:
  • Delivery state and proof:

Handoff

Next action

  • Exact next action:
  • Changed files / branch / active processes:
  • State updated in PROJECT_STATE.md:
  • Owner rules stated in the required reads:
  • Defects this cycle leaves unfixed (all go to the next cycle's Issue):
  • At most one next candidate and why (after evidence):