Spec Kitty

Work Packages: Nightly census re-pin and timeout headroom

Mission nightly-census-and-timeout-headroom-01M3PM2S — #5367, #5378. Planned under planner-priti.

Subtask Format: [Txxx] [P?] Description

Path Conventions

Repo-root-relative paths; no kitty-specs/ paths in code_change WPs.

Work Package WP01: Census honesty — #5367 (Priority: P1) 🎯 MVP

Goal: SC-011 census counts only genuine raw material; test green. Independent Test: pytest tests/specify_cli/bulk_edit/test_occurrence_map_field_paths.py RED (26≠19) → GREEN. Prompt: tasks/WP01-census-honesty.md

Included Subtasks

• □ T001 Record red-first evidence: census test fails 26 == 19 on the branch base.
• □ T002 Remove the references: block from packs/built-in/toolguides/{java,javascript,python}-supply-chain.toolguide.yaml.
• □ T003 spec-kitty doctrine regenerate-graph; confirm only pack-manifest.yaml hashes change (no *.graph.yaml).
• □ T004 Re-pin test_governance_occurrences_and_files_match_sc011: len(raw) == 20, add common-docs to raw_files and raw_files & migrate_files, provenance note naming 3e09226f/#5324; fix stale "5 of 7"/"7 RAW files" docstring counts touched by this change.

Implementation Notes

The removed refs are inert (extractor path-only resolver). Do not add outbound edges (design decision DD-1).

Parallel Opportunities

Independent of WP02.

Dependencies

None.

Risks & Mitigations

Any *.graph.yaml diff after regeneration ⇒ refs were not inert; stop.

Work Package WP02: Nightly timeout headroom — #5378 (Priority: P1)

Goal: every capped interpreter shard and the out-of-matrix leg has ≥1.5× measured headroom. Independent Test: new tests/ci/test_nightly_timeout_headroom.py RED against current caps → GREEN. Prompt: tasks/WP02-nightly-timeout-headroom.md

Included Subtasks

• □ T005 Add tests/ci/test_nightly_timeout_headroom.py: parse ci-nightly.yml only; every interpreter shard (roster job keys) and specify-cli-out-of-matrix must carry a structured # headroom: comment and cap ≥ ceil((max + 0:30) × 1.5); run it RED.
• □ T006 Raise caps in .github/workflows/ci-nightly.yml: shard 1 → 20, shard 2 → 31, shard 3 → 35, shard 5 → 35, out-of-matrix → 69 (shards 4, 6 keep 25); record the derivation as # headroom: comments (censored maxima as >=); update _interpreter_shard_roster.py docstring to point there; CHANGELOG entry for #5367/#5378.
• □ T007 Run test_interpreter_shard_coverage.py, test_module_shard_registry.py, tests/ci/test_nightly_exit_code_honesty.py, tests/ci/test_interpreter_matrix_env_pinning.py, the new test.

Implementation Notes

Only timeout-minutes: values and comments change; suite run lines untouched.

Parallel Opportunities

Independent of WP01.

Dependencies

None.

Risks & Mitigations

Workflow-shape tests parse ci-nightly.yml; keep job keys and run lines unchanged.

Dependency & Execution Summary

• Sequence: WP01 ∥ WP02 (no dependencies).
• Parallelization: fully parallel; disjoint files.
• MVP Scope: WP01 (only test red on the nightly).

Requirements Coverage Summary

Requirement ID	Covered By Work Package(s)
FR-001	WP01
FR-002	WP01
FR-003	WP02
FR-004	WP02
NFR-001	WP01
NFR-002	WP01, WP02
C-001	WP02
C-002	WP02
C-003	WP01, WP02

Subtask Index (Reference)

Subtask ID	Summary	Work Package	Priority	Parallel?
T001	Red-first census evidence	WP01	P1	No
T002	Remove toolguide refs	WP01	P1	No
T003	Regenerate graph/manifest	WP01	P1	No
T004	Re-pin census 19→20	WP01	P1	No
T005	Headroom test (red)	WP02	P1	Yes
T006	Raise caps	WP02	P1	No
T007	Gate files	WP02	P1	No