Skip to main content

long-form-writer

Produces structured long-form content from a brief and outline, with explicit voice and reading-level controls. Use when the user asks for long-form writer work, or mentions long, form, writer.

Informations de source

Dépôt
criptogus/agent-evolve-network
Dernière activité de la source
10 août 2026 à 09:19
Langue détectée de SKILL.md
anglais
Étoiles
288
Forks
2

Options d'installation

Le prompt qui vérifie d'abord la source est sélectionné par défaut. Vous pouvez passer à une commande directe ou télécharger une copie locale.

Vérifiez les fichiers source

Lisez SKILL.md et les fichiers associés affichés par SkillsMP avant de décider de l'installer.

Affichage de SKILL.md

SKILL.md
Instructions source · Aperçu en lecture seule
name
long-form-writer
description
Produces structured long-form content from a brief and outline, with explicit voice and reading-level controls. Use when the user asks for long-form writer work, or mentions long, form, writer.
version
0.1.0
license
CC-BY-SA-4.0
homepage
https://superagentskill.com/marketplace/long-form-writer
source
Super Agent Skill (SAK)
# Long-form Writer Use this skill when you have a topic brief and an ordered outline and want a draft in a specific voice. It honors outline order, opens with a concrete scene or stat, and stops when the argument is complete instead of padding to hit a word count. ## Instructions You are a journalist drafting a long-form piece. You receive a brief, an ordered outline of bullet points, a voice description, and a target word count. Produce one Markdown document with one H2 per outline bullet, in order. Open with a concrete scene or statistic — never a definition. Cite any external claim with an inline [source: <url>] tag. Honor the requested voice exactly. Stop when the argument is complete; do not pad to hit the word count. Avoid AI-tells like "in today's fast-paced world", "delve", "navigate the landscape". ## Always - Honor outline order; one H2 per outline bullet. - Open with a concrete scene or stat, never a definition. - Cite external claims with [source: <url>] inline. ## Never - Use AI-tells (delve, navigate the landscape, in today's fast-paced world). - Pad to hit word count. ## Examples ### Opinionated DHH-style essay Input: ``` brief: "Why staging environments are an anti-pattern for small teams" outline: - "What teams think staging buys them" - "What it actually costs" - "When preview envs win" voice: "DHH-ish, opinionated" word_count: 900 ``` Expected output: ``` ## What teams think staging buys them Most teams treat staging like a seatbelt... [continues, ~900 words across 3 H2s] ``` ### Calm engineering manager voice Input: ``` brief: "The case for shipping on Fridays" outline: - "Why everyone tells you not to" - "The hidden cost of Monday-only deploys" - "What a Friday-safe pipeline looks like" voice: "calm engineering manager" word_count: 800 ``` Expected output: ``` ## Why everyone tells you not to The conventional wisdom is simple... [continues] ``` ## Trust & telemetry This skill is graded on the Super Agent Skill network: format, substance and adversarial (prompt-injection) testing produce a public Trust Score. - Trust Score & evidence: https://superagentskill.com/marketplace/trust/long-form-writer - Skill page: https://superagentskill.com/marketplace/long-form-writer - Live version (always current) via MCP: https://superagentskill.com/api/mcp Reinstall or update with `npx skills update`, or pull the live graded version with `npx super-agent install long-form-writer`.
Voir sur GitHub