outline-refiner

Name: outline-refiner
Author: WILLOSCAR

489stars39forks

How to install outline-refiner?

willoscar/research-units-pipeline-skills/outline-refiner

$npx -y skills add WILLOSCAR/research-units-pipeline-skills --skill outline-refiner

Installs into the current project.

Summary

Planner-pass coverage + redundancy report for an outline+mapping, producing outline/coverage_report.md and outline/outline_state.jsonl. Trigger: planner, dynamic outline, outline refinement, coverage report, 大纲迭代, 覆盖率报告. Use when: you have outline/outline.yml + `out

Files · 1

View on GitHub

SKILL.md

1	# Outline Refiner (Planner pass, NO PROSE)
2
3	Goal: make the outline auditable by adding an explicit planner stage that answers:
4	- Do we have enough mapped evidence per H3?
5	- Are the same few papers reused everywhere?
6	- Are subsection axes still generic/scaffold-y?
7	- Is the outline likely to produce a paper-like structure (final ToC budget: ~6–8 H2; fewer, thicker H3s)?
8
9	This is a deterministic “planner” unit: it must not write survey prose.
10
11	## Inputs
12
13	Required:
14	- `outline/outline.yml`
15	- `outline/mapping.tsv`
16
17	Optional (best-effort diagnosis; may be missing early in the pipeline):
18	- `outline/OUTLINE_BUDGET_REPORT.md` (if present: explains recent merges; helps interpret mapping/coverage changes)
19	- `papers/paper_notes.jsonl` (for evidence levels)
20	- `outline/subsection_briefs.jsonl` (for axis specificity)
21	- `GOAL.md` (for scope drift hints)
22
23	## Outputs
24
25	- `outline/coverage_report.md` (bullets + small tables; NO PROSE)
26	- `outline/outline_state.jsonl` (append-only JSONL; one record per run)
27
28	## Workflow (planner pass, NO PROSE)
29
30	1. Parse `outline/outline.yml` to enumerate H2 sections + H3 subsections (section sizing / budget).
31	- If `outline/OUTLINE_BUDGET_REPORT.md` exists, use it as the merge/change log so the coverage report can explain why structure changed.
32	2. Read `outline/mapping.tsv` and compute per-H3 coverage and reuse hotspots.
33	3. If `papers/paper_notes.jsonl` exists, summarize evidence levels (fulltext/abstract/title) for mapped papers.
34	4. If `outline/subsection_briefs.jsonl` exists, compute axis specificity (generic vs specific axes) per H3.
35	5. Optionally use `GOAL.md` to flag obvious scope drift (keywords not reflected in outline).
36	6. Write `outline/coverage_report.md` and append a run record to `outline/outline_state.jsonl`.
37
38	## Freeze policy
39
40	- If `outline/coverage_report.refined.ok` exists, the script will not overwrite `outline/coverage_report.md`.
41
42	## Script
43
44	### Quick Start
45
46	- `uv run python .codex/skills/outline-refiner/scripts/run.py --help`
47	- `uv run python .codex/skills/outline-refiner/scripts/run.py --workspace <workspace>`
48
49	### All Options
50
51	- `--workspace <dir>`: workspace root
52	- `--unit-id <U###>`: unit id (optional; for logs)
53	- `--inputs <semicolon-separated>`: override inputs (rare; prefer defaults)
54	- `--outputs <semicolon-separated>`: override outputs (rare; prefer defaults)
55	- `--checkpoint <C#>`: checkpoint id (optional; for logs)
56
57	### Examples
58
59	- Planner-pass diagnostics after `section-mapper`:
60	- `uv run python .codex/skills/outline-refiner/scripts/run.py --workspace <workspace>`
61
62	## Troubleshooting
63
64	### Issue: report is missing evidence-level or axis-specificity columns
65
66	Cause:
67	- Optional inputs are missing (no `papers/paper_notes.jsonl` and/or no `outline/subsection_briefs.jsonl`).
68
69	Fix:
70	- Run `paper-notes` and/or `subsection-briefs`, then rerun `outline-refiner`.

People who liked this also viewed

6 picks

Type

outline-refiner

489stars39forks

View on GitHub

How to install outline-refiner?

willoscar/research-units-pipeline-skills/outline-refiner

$npx -y skills add WILLOSCAR/research-units-pipeline-skills --skill outline-refiner

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Outline Refiner (Planner pass, NO PROSE)
2
3	Goal: make the outline auditable by adding an explicit planner stage that answers:
4	- Do we have enough mapped evidence per H3?
5	- Are the same few papers reused everywhere?
6	- Are subsection axes still generic/scaffold-y?
7	- Is the outline likely to produce a paper-like structure (final ToC budget: ~6–8 H2; fewer, thicker H3s)?
8
9	This is a deterministic “planner” unit: it must not write survey prose.
10
11	## Inputs
12
13	Required:
14	- `outline/outline.yml`
15	- `outline/mapping.tsv`
16
17	Optional (best-effort diagnosis; may be missing early in the pipeline):
18	- `outline/OUTLINE_BUDGET_REPORT.md` (if present: explains recent merges; helps interpret mapping/coverage changes)
19	- `papers/paper_notes.jsonl` (for evidence levels)
20	- `outline/subsection_briefs.jsonl` (for axis specificity)
21	- `GOAL.md` (for scope drift hints)
22
23	## Outputs
24
25	- `outline/coverage_report.md` (bullets + small tables; NO PROSE)
26	- `outline/outline_state.jsonl` (append-only JSONL; one record per run)
27
28	## Workflow (planner pass, NO PROSE)
29
30	1. Parse `outline/outline.yml` to enumerate H2 sections + H3 subsections (section sizing / budget).
31	- If `outline/OUTLINE_BUDGET_REPORT.md` exists, use it as the merge/change log so the coverage report can explain why structure changed.
32	2. Read `outline/mapping.tsv` and compute per-H3 coverage and reuse hotspots.
33	3. If `papers/paper_notes.jsonl` exists, summarize evidence levels (fulltext/abstract/title) for mapped papers.
34	4. If `outline/subsection_briefs.jsonl` exists, compute axis specificity (generic vs specific axes) per H3.
35	5. Optionally use `GOAL.md` to flag obvious scope drift (keywords not reflected in outline).
36	6. Write `outline/coverage_report.md` and append a run record to `outline/outline_state.jsonl`.
37
38	## Freeze policy
39
40	- If `outline/coverage_report.refined.ok` exists, the script will not overwrite `outline/coverage_report.md`.
41
42	## Script
43
44	### Quick Start
45
46	- `uv run python .codex/skills/outline-refiner/scripts/run.py --help`
47	- `uv run python .codex/skills/outline-refiner/scripts/run.py --workspace <workspace>`
48
49	### All Options
50
51	- `--workspace <dir>`: workspace root
52	- `--unit-id <U###>`: unit id (optional; for logs)
53	- `--inputs <semicolon-separated>`: override inputs (rare; prefer defaults)
54	- `--outputs <semicolon-separated>`: override outputs (rare; prefer defaults)
55	- `--checkpoint <C#>`: checkpoint id (optional; for logs)
56
57	### Examples
58
59	- Planner-pass diagnostics after `section-mapper`:
60	- `uv run python .codex/skills/outline-refiner/scripts/run.py --workspace <workspace>`
61
62	## Troubleshooting
63
64	### Issue: report is missing evidence-level or axis-specificity columns
65
66	Cause:
67	- Optional inputs are missing (no `papers/paper_notes.jsonl` and/or no `outline/subsection_briefs.jsonl`).
68
69	Fix:
70	- Run `paper-notes` and/or `subsection-briefs`, then rerun `outline-refiner`.

People who liked this also viewed

6 picks

Type