QRSPI Implement

"Amateurs practice until they get it right. Professionals practice until they can't get it wrong."

Core Philosophy

The Implement phase executes an already-approved plan mechanically -- every design decision was settled by Spec and Plan. Each plan phase is one vertical slice, and each slice runs the full Red-Green-Refactor loop end-to-end before the next begins. The discipline is test-first and per-slice: a slice is not done until its RED proof and GREEN proof are recorded on disk. One slice is ideally one fresh session -- the per-slice log is the resumption point, so context never has to span the whole feature.

Non-Negotiable Constraints:

1. APPROVED-GATED -- never start without plan.md status: approved; never implement from memory
2. BASELINE GREEN FIRST -- run the test suite before any change; a red baseline is a STOP, not yours to fix
3. ONE SLICE AT A TIME -- execute plan phases in order; never reorder, parallelize, or skip
4. TEST FIRST OR STOP -- production code written before a failing test is a STOP condition
5. RGR PER SLICE -- RED (test fails) -> GREEN (minimal code, build+test pass) -> REFACTOR (stays green)
6. RECORD THE PROOF -- write RED and GREEN command output to implementation/slice-NN-{name}.md
7. INVENTION = STOP -- anything not in the plan stops the slice; re-run /qrspi-plan, do not improvise
8. CONTEXT BUDGET: keep utilization under 40%. At 60% (or end of a slice), checkpoint to the slice log and tell the user to start a fresh session for the next slice.

Workflow

PRE-FLIGHT
    [ ] Locate the feature folder thoughts/shared/qrspi/YYYY-MM-DD-{slug}/
    [ ] Read plan.md. If absent -> STOP; route the user to /qrspi-plan
    [ ] Confirm plan.md status: approved. If only ready-for-review -> STOP; ask the human to approve
    [ ] Run the baseline test suite. If red -> STOP; report; do not fix the baseline
    [ ] Read the ENTIRE plan; identify the next unfinished slice from implementation/

SLICE LOOP  (one slice = one plan phase, ideally one fresh session)
    1. RED   — write the failing test from the phase; RUN it; it MUST FAIL.
               If it passes, the test is wrong -> fix the test, not the code.
    2. GREEN — write the minimal production code; RUN build + tests; they MUST PASS.
    3. REFACTOR — clean up; RUN again; it MUST stay GREEN.
    4. RECORD — write implementation/slice-NN-{name}.md: RED output, GREEN output, files changed.
    5. CHECKPOINT — suggest one commit for the slice; if context > 40% or the slice is done,
                    tell the user to start a fresh session for the next slice.

REPORT
    Slices complete / total · last verification result · next slice (or "feature complete").

Exit criteria: each executed slice has a slice-NN-{name}.md log holding its RED proof and GREEN proof; the build/tests are GREEN; a per-slice commit was suggested; the user knows the next slice or that the feature is complete.

State Block

<qrspi-implement-state>
phase: PRE-FLIGHT | RED | GREEN | REFACTOR | RECORD | CHECKPOINT | COMPLETE
feature_folder: thoughts/shared/qrspi/YYYY-MM-DD-{slug}/
plan_present: true | false          # MUST be true to proceed
plan_approved: true | false         # MUST be true to proceed
baseline_green: true | false        # MUST be true before any change
current_slice: [NN]
slices_total: [count]
slices_complete: [count]
last_verification: red | green | pending   # RED step expects red; GREEN/REFACTOR expect green
context_budget: under-40 | approaching-60 | checkpoint-now
</qrspi-implement-state>

Output Template

See references/slice-log-template.md for the per-slice implementation/slice-NN-{name}.md structure -- the RED proof, GREEN proof, files-changed list, and the resume note.

Integration with Other Skills