Files
claw/docs/superpowers/plans/2026-04-20-scene-skill-102-pseudoprod-batch-selection-plan.md

128 lines
3.7 KiB
Markdown

# Scene Skill 102 Pseudo-Production Batch Selection Plan
> Date: 2026-04-20
> Status: Draft
> Upstream Design: `docs/superpowers/specs/2026-04-20-scene-skill-102-pseudoprod-batch-selection-design.md`
## Plan Intent
Select the first pseudo-production validation batch from the 102 final materialized skills.
This plan is selection-only. It does not run pseudo-production execution.
## Fixed Inputs
1. `examples/scene_skill_102_final_materialization_2026-04-19/scene_skill_102_index.json`
2. `tests/fixtures/generated_scene/scene_skill_102_static_validation_2026-04-20.json`
3. `tests/fixtures/generated_scene/scene_skill_102_dispatch_dry_run_validation_2026-04-20.json`
4. `tests/fixtures/generated_scene/scene_skill_102_full_direct_mock_execution_2026-04-20.json`
5. `tests/fixtures/generated_scene/scene_skill_102_pseudoprod_readiness_2026-04-20.json`
## Planned Outputs
1. `tests/fixtures/generated_scene/scene_skill_102_pseudoprod_batch_selection_2026-04-20.json`
2. `docs/superpowers/reports/2026-04-20-scene-skill-102-pseudoprod-batch-selection-report.md`
## Allowed Files
1. the planned output JSON
2. the planned report
## Forbidden Files
1. `src/generated_scene/analyzer.rs`
2. `src/generated_scene/generator.rs`
3. `src/generated_scene/ir.rs`
4. `examples/scene_skill_102_final_materialization_2026-04-19/skills/**`
5. `tests/fixtures/generated_scene/scene_execution_board_2026-04-18.json`
## Phase 0: Freeze Selection Boundary
### Tasks
1. Confirm all 102 scenes are direct-mock-pass.
2. Confirm this plan does not execute browser automation or real network access.
3. Confirm this plan does not update official board status.
### Acceptance Criteria
1. Selection starts from a clean `102 / 102` local mock baseline.
2. Selection does not mutate generated skills or runtime code.
## Phase 1: Build Eligible Candidate Set
### Tasks
1. Read pseudo-production readiness records.
2. Keep only `pseudo-prod-ready` scenes.
3. Exclude `real-env-required` scenes from first batch.
4. Join with direct mock results and static/dispatch readiness.
### Acceptance Criteria
1. Every selected candidate is static validated.
2. Every selected candidate is dispatch ready.
3. Every selected candidate is direct-mock-pass.
## Phase 2: Select Balanced First Batch
### Tasks
Select `10` scenes with archetype balance:
1. `paginated_enrichment`: 4
2. `multi_mode_request`: 2
3. `single_request_enrichment`: 2
4. `multi_endpoint_inventory`: 1
5. `page_state_eval`: 1
### Acceptance Criteria
1. The selected batch contains exactly `10` scenes.
2. The batch excludes host-bridge and local-doc runtime-dependent scenes.
3. Every selected scene has a deterministic invocation input.
## Phase 3: Define Evidence Checklist
### Tasks
For each selected scene, define required evidence:
1. deterministic invocation input
2. console log
3. network log or request summary
4. screenshot if browser target page is required
5. exported file if produced
6. generation report path
7. failure taxonomy slot
### Acceptance Criteria
1. Every selected scene has a complete checklist.
2. Checklist does not require production credentials to be stored in the repository.
## Phase 4: Publish Selection Report
### Tasks
1. Write selection JSON.
2. Write selection report.
3. Summarize selected and deferred scenes.
### Acceptance Criteria
1. The report states this is selection-only.
2. The report does not claim pseudo-production execution.
3. The report identifies the next bounded execution plan.
## Completion Criteria
This plan is complete when the first pseudo-production batch selection JSON and report are published.
## Stop Statement
Stop after publishing selection assets.
Do not execute pseudo-production validation under this plan.