TypingMindTypingMindThe best frontend for AI
Subsection Writer logo

Subsection Writer

Community
WILLOSCAR
subsection-writer

Write survey prose into per-section files under `sections/` so each unit can be QA'd independently before merging.

Overview

PublisherWILLOSCAR
Repositoryresearch-units-pipeline-skills
Skill namesubsection-writer
Stars
513
Forks
39
Bundled files
14
Links
Skill sourceRepository
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 14 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by WILLOSCAR on GitHub. Read the source before you install it.

Installation

Install the Subsection Writer AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/WILLOSCAR/research-units-pipeline-skills/tree/main/.codex/skills/subsection-writer
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/WILLOSCAR/research-units-pipeline-skills.git /tmp/research-units-pipeline-skills
mkdir -p .claude/skills
cp -r /tmp/research-units-pipeline-skills/.codex/skills/subsection-writer .claude/skills/subsection-writer
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Subsection Writer in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Subsection Writer on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Subsection Writer is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Subsection Writer (compatibility router)

Triggers & routing

  • Trigger: subsection writer, per-section writing, split sections, sections/, 分小节写, 按章节拆分写作.
  • Use when: Approve C2 is recorded and writer packs exist (outline/writer_context_packs.jsonl); you want evidence-bounded drafting without a monolithic one-shot draft.

Purpose: write or refine per-section survey prose under sections/ while keeping the current pipeline contract unchanged.

Compatibility mode:

  • output paths stay the same
  • scripts/run.py still handles approval checks, missing-file bootstrap, and sections/sections_manifest.jsonl
  • writing guidance now lives in references/ instead of being encoded primarily in the script

Load Order

Always read:

  • references/overview.md
  • references/paragraph_jobs.md
  • references/paragraph_job_archetypes.md when adjusting paragraph-role behavior or refactoring writer policy out of Python
  • references/bootstrap_assembly.md when reasoning about compatibility-mode bootstrap behavior

Read as needed:

  • references/opener_catalog.md when paragraph 1 sounds generic or narrated
  • references/contrast_moves.md when building A-vs-B comparison paragraphs
  • references/eval_anchor_patterns.md when making performance / robustness / benchmark claims
  • references/limitation_moves.md when adding caveats or local conclusions
  • references/examples_good.md and references/examples_bad.md for calibration only

Machine-readable contract:

  • assets/subsection_writer_context.schema.json
  • assets/bootstrap_paragraph_templates.json
  • assets/paragraph_job_templates.json

Inputs

Required:

  • DECISIONS.md (must include Approve C2)
  • outline/outline.yml
  • outline/writer_context_packs.jsonl (preferred)
  • citations/ref.bib

Optional but useful:

  • outline/subsection_briefs.jsonl
  • outline/evidence_drafts.jsonl
  • outline/evidence_bindings.jsonl
  • outline/anchor_sheet.jsonl
  • outline/chapter_briefs.jsonl

Outputs

Keep the current contract:

  • sections/S<sub_id>.md for H3 bodies
  • sections/sections_manifest.jsonl, refreshed across all section files already produced by the three writer Skills
  • sections/h3_bodies.refined.ok after a model or human has reviewed the generated H3 bodies and is submitting U100 for mandatory acceptance

front-matter-writer owns Abstract, Introduction, Related Work, Discussion, and Conclusion. chapter-lead-writer owns H2 lead blocks. This Skill reads those files while refreshing the shared manifest but does not claim their authorship.

Writer policy

The active rule is move coverage, not paragraph quota. A subsection should cover the necessary argument moves the pack supports; do not pad to a fixed count when the evidence does not justify it.

Opener / ending policy:

  • generate 2-4 opener candidates from the pack's actual tension, contrast, protocol, or limitation signals; keep the most content-bearing option instead of reusing one stock stem everywhere
  • let subsection endings emerge from evidence-bearing comparison / limitation material; do not append a fixed “safest synthesis / decision rule” closer just to make the paragraph feel finished
  • normalize internal axis labels into natural reader prose; slash-style brief handles should remain planning metadata, not leak unchanged into the paper

Script boundary

Use scripts/run.py as a helper only:

  • it may bootstrap missing H3 files and refresh the manifest
  • bootstrap prose is assembled deterministically from writer packs and the versioned template assets; it is inspectable fallback material, not a claim of model authorship or final prose quality
  • it must not be treated as the canonical source of prose shape or voice policy
  • it must not self-certify its bootstrap prose: the script never creates sections/h3_bodies.refined.ok
  • the marker is only U100's submission attestation; it is created before the downstream writer-selfloop, so it cannot mean that the self-loop already passed
  • mandatory acceptance independently compares H3 sentences with the Run-selected writer template assets and rejects literal residue above the Pipeline limit
  • writer-selfloop calls the shared strict section checker, which recomputes the H3 measure before the report can PASS
  • the mandatory pipeline-auditor measures the entire merged draft and writes a scorecard whose verdict and dimensions the Harness projects into the evaluation ledger
  • the current 10% limit is an initial policy target; a published current-contract replay measures 0/226 residue for one retained Artifact set, while clean from-scratch and cross-topic reproduction remain open
  • if the marker predates a writer input or the writer script, it is stale and the next run removes it before regenerating bootstrap prose

Quick Start

  • uv run python .codex/skills/subsection-writer/scripts/run.py --workspace <workspace>

Troubleshooting

  • If the pack is thin, stop and route upstream instead of padding prose.
  • If the subsection sounds narrated, reload references/opener_catalog.md and references/examples_bad.md.
  • If a claim lacks protocol context, reload references/eval_anchor_patterns.md before rewriting.

Execution notes

When running in compatibility mode, scripts/run.py currently consumes:

  • DECISIONS.md for Approve C2
  • outline/outline.yml to enumerate chapter / subsection files
  • outline/writer_context_packs.jsonl as the primary drafting input
  • citations/ref.bib for in-scope citations
  • outline/subsection_briefs.jsonl, outline/evidence_drafts.jsonl, outline/evidence_bindings.jsonl, outline/anchor_sheet.jsonl, and outline/chapter_briefs.jsonl as optional enrichment sources

Script

Quick Start

  • uv run python .codex/skills/subsection-writer/scripts/run.py --workspace <workspace>

All Options

  • --workspace <dir>
  • --unit-id <id>
  • --inputs <a;b;...>
  • --outputs <a;b;...>
  • --checkpoint <C*>

Examples

  • uv run python .codex/skills/subsection-writer/scripts/run.py --workspace <workspace>

Troubleshooting

  • If DECISIONS.md lacks Approve C2, stop and fix approval first.
  • If outline/writer_context_packs.jsonl is thin, reroute upstream instead of padding prose.
  • If citation scope looks wrong, re-check citations/ref.bib and the writer packs before editing output text.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Subsection Writer AI skill do?

Write survey prose into per-section files under `sections/` so each unit can be QA'd independently before merging.

Why use Subsection Writer on TypingMind?

Because you install it once and use it with any model. Subsection Writer is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Subsection Writer in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/WILLOSCAR/research-units-pipeline-skills/tree/main/.codex/skills/subsection-writer. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Subsection Writer?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Subsection Writer?

As many as you like. As long as a model supports skills, you can use Subsection Writer with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Subsection Writer AI skill free?

It is published on GitHub by WILLOSCAR. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

Related AI Skills

View all
openclaw logo

Release Openclaw Mac

openclaw
OrganizationPopular

Run or recover OpenClaw macOS release signing, notarization, appcast, and asset promotion.

82.2K
391K
Instructions
View
openclaw logo

Release Openclaw Maintainer

openclaw
OrganizationPopular

Prepare, publish, recover, or verify OpenClaw beta, stable, and extended-stable releases, including approved backports.

82.2K
391K
11 files
View
openclaw logo

Release Openclaw Plugin Testing

openclaw
OrganizationPopular

Plan and run pre-release OpenClaw plugin validation across bundled plugins, package artifacts, lifecycle commands, doctor/fix, config round-trip, gateway startup, SDK compatibility, Docker E2E,…

82.2K
391K
1 files
View
openclaw logo

Solid

openclaw
OrganizationPopular

Write, port, review, or test Solid 2 UI in OpenClaw (Control UI, plugin views, Canvas hosts), including state ownership, Lit interop during the migration, native overlays, testing, performance, and…

82.4K
391.7K
9 files
View

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇