Open-source screenwriting knowledge infrastructure for writers and agents.
Route, generate, review, and orchestrate narrative, branded, and interactive scripts.
Not a prompt dump. Not a single-method gospel. Not a UI-first product. Durable creative infrastructure for screenplay work: routable knowledge, clear workflow contracts, reusable review logic, and community-driven correction loops.
Request
Turn this idea into a feature-film beat sheet:
"A journalist who has spent years avoiding the truth behind her father's death
is forced back to her mining hometown to investigate an old case."
Selected route → skill.structure-beat + wp.structure-beat-outline
Artifact excerpt
## Beat List
- Opening imbalance: She avoids every mining story that crosses her desk.
- Lock-in: A fragment from her father's case file forces her back home.
- Midpoint turn: She learns her own silence helped protect the cover-up.
Full chain: request · artifact · more examples
This repository isn't a prompt library or a single-method writing course. It's durable, composable screenwriting infrastructure built for both humans and LLM agents.
It classifies your request by intent, medium, stage, and desired output, then loads only the relevant knowledge — protocol, rubric, and reference atoms — to produce a structured artifact. It self-checks before handing you the result. When you push back, your feedback becomes better assets for the next run.
If you're a writer: it's a structured development and review toolkit. If you're building agent workflows: it's explicit routing logic with bounded loading and machine-readable contracts.
flowchart LR
S["Preflight Sync<br/>upstream SHA check"] --> A["Request"]
A --> B["Classify & Route<br/>intent × medium × stage × output"]
B --> C["Load Bundle<br/>protocol + rubric + atoms"]
C --> D["Generate Artifact"]
D --> E["Self-Check<br/>rubric-based quality gate"]
E --> F["Human Feedback"]
F --> G["Improve Assets"]
G -.->|"next run"| S
git clone https://github.com/XucroYuri/how-to-make-script.git ~/.local/share/how-to-make-script
# Updates: git -C ~/.local/share/how-to-make-script pull --ff-onlyClaude Code
mkdir -p ~/.claude/skills
ln -sfn ~/.local/share/how-to-make-script ~/.claude/skills/how-to-make-scriptCodex
[[skills.config]]
path = "/Users/<you>/.local/share/how-to-make-script"
enabled = trueOther platforms (Gemini CLI, OpenCode, OpenClaw): same pattern — clone locally, point your tool at the checkout. See SKILL.md for the root entrypoint.
| Surface | Count |
|---|---|
| Sub-skills | 29 in skills/ |
| Output types | 30 in supported-outputs.md |
| Knowledge atoms | 114 in knowledge/ |
| Workflow protocols | 33 in knowledge/20-workflows/ |
| Evaluation rubrics | 31 in knowledge/60-rubrics/ |
| Route fixtures | 119 in fixtures.json |
| Knowledge .md files | 189 in knowledge/ |
| Validation scripts | 19 in scripts/ |
| Test modules | 19 in tests/ |
| Golden examples | 10 flows in examples/golden/ |
| Reference packs | 10 in examples/reference-packs/ |
| Bilingual docs | 17 pairs in docs/ |
Writing & development — logline, premise, beat sheet, outline, scene draft, screenplay draft, dialogue polish
Review & correction — rewrite diagnosis, quality gates, scope correction, boundary maps
Platform-specific — commercial scripts, branded film scripts, interactive branch maps
Expression & downstream — voice style guides, visual language packs, screenplay-to-video briefs
Team & system — team workflow blueprints, expert subagent casting, handoff protocols
For writers — golden examples · supported outputs · narrative reference pack · adaptive quality checking
For agent builders — SKILL.md (root orchestration) · router matrix · routing policy · context loading policy · content model
For contributors — contributing guide · support ladder · changelog
This project grows through high-signal disagreement.
| Channel | Use for |
|---|---|
| Discussions | Questions, rebuttals, field notes, rival paths |
| Issues | Concrete route, rubric, or asset changes |
| Security | Private vulnerability reporting |
Current: Usable research-first screenplay monorepo. Strong routing, loading, and quality-gating infrastructure. Broad narrative/commercial/interactive coverage.
Open gaps: Runtime execution not yet implemented. Bundle-planner enforcement incomplete. Edge-case fixture depth uneven. Some genre/stage-specific knowledge still thin. Discussion-to-asset conversion remains manual.
Good first contributions: challenge an overly broad claim, add a counterexample, improve a doc path, or turn a route mismatch into a fixture.
Contributing · Code of Conduct · Support · Security · Citation · License