# Spine Framework

> Structures long-form technical articles with the SPINE narrative framework for clarity and retention. Use this skill when writing or restructuring essays, deep-dives, or tutorial articles. Do not use when/for short README/API docs (use documentation-standards) or Agent Skill authoring (use skill-authoring).

- Skill: `jagreehal/spine-framework` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add jagreehal/spine-framework`
- Raw SKILL.md: https://api.skillmd.com/api/skills/jagreehal/spine-framework/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: jagreehal (https://skillmd.com/u/jagreehal)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/jagreehal/spine-framework

---


# SPINE Framework

## Critical rules

- Substance first: **no insight, no article.** Point outranks polish.
- SPINE is an evaluation lens — never use S/P/I/N/E as article headings; use information-carrying titles.
- Prove claims with running code or cited trusted sources. Never fabricate output or stats.
- Admit limits (Nuance). Land the plane (Exit) with insight + callback — not a recap or fluff ending.
- Before drafting or scoring, read [references/elements.md](references/elements.md).

## Workflow

1. State the one-sentence Point (no "how"). If empty, stop — no article yet.
2. Draft Stakes in the first 30 seconds (who, felt problem, what they'll do, scope).
3. Build Illustration that actually proves the Point; run every code sample.
4. Add Nuance: ≥2 failure modes, costs, when NOT to use.
5. Write Exit: restate insight, when to use, one takeaway, opening callback.
6. Pass Voice and Clean last. Run the pre-flight in [references/scoring.md](references/scoring.md) and score.

## Resources

- [references/elements.md](references/elements.md) — Stakes/Point/Illustration/Nuance/Exit/Voice/Clean detail. Read when writing or reviewing.
- [references/scoring.md](references/scoring.md) — pre-flight checklist and score bands. Read before publishing.

## Validation

- [ ] One-sentence insight in first 20%; repeated at end
- [ ] Reader can decide relevance in ~20 seconds
- [ ] All code run; claims sourced; no fabricated output
- [ ] Failure modes and when-not-to-use present
- [ ] Exit restates insight with callback; no fluff ending
- [ ] No SPINE-letter headings; pre-flight checklist passed

## Constraints

- Not for README/API reference (`documentation-standards`) or voice-preserving line edits alone (`structured-writing`).
- Related: `documentation-standards`, `structured-writing`, `concise-output`.

