Long-Form Drafter

Long form

by timscheuerai · MIT license · ★ 37 on GitHub

Long form cover

Draft written long-form content (newsletter issue, blog post, Substack issue) in your voice. 800-2000 words, structured intro + body + CTA, paragraph-by-paragraph not list-by-list.

---
name: long-form
description: Draft written long-form content (newsletter issue, blog post, Substack issue) in your voice. 800-2000 words, structured intro + body + CTA, paragraph-by-paragraph not list-by-list. Pulls voice from /linkedin-copywriter's corpus (same author, same register · longer arc). Runs the brand-doc skill checklist before output. Logs the row in the Notion Content DB and (for newsletters) drafts a separate punchy subject line. Use when asked to "write a newsletter", "draft a blog post", "long-form X", "Substack issue about Y", "weekly newsletter", or "/long-form". For spoken video scripts (YouTube longs / shorts) use /youtube-script instead · the cadence is fundamentally different.
---

# long-form

Written long-form content. One author voice (yours), three default
shapes (newsletter / blog / Substack issue). Same skill, different
register and length per shape.

This is **written long-form**. For **spoken long-form** (YouTube
longs, podcast monologues, talking-head video scripts), use
`/youtube-script` · spoken cadence is structurally different and
trying to merge them produces worse output for both.

## When to use

Overview

Written long-form content. One author voice (yours), three default shapes (newsletter / blog / Substack issue). Same skill, different register and length per shape.

Manual install (npx, ZIP)
Claude Codenpx skills add timscheuerai/content-vault --skill long-form -g -a claude-code
Claude Code (degit)npx degit timscheuerai/content-vault/skills/long-form#main ~/.claude/skills/long-form
Claude.ai / desktopZip the folder and upload it under Settings → Capabilities → Skills.
Codex, Cursor, other agentsPoint them at SKILL.md. The other files are plain text.

Connects with Agent Alley MCP · agentalley.io/mcp

Files included

  • SKILL.md10.1 KB
Read SKILL.md here
timscheuerai/main1 file shown
SKILL.md
Show the full text260 lines

long-form

Written long-form content. One author voice (yours), three default shapes (newsletter / blog / Substack issue). Same skill, different register and length per shape.

This is written long-form. For spoken long-form (YouTube longs, podcast monologues, talking-head video scripts), use /youtube-script · spoken cadence is structurally different and trying to merge them produces worse output for both.

When to use

Trigger on:

  • "Write a newsletter issue"
  • "Draft a blog post about X"
  • "Long-form / Substack issue on Y"
  • "Weekly newsletter"
  • "Article for the website"
  • User explicitly types /long-form

Skip for:

  • LinkedIn / X posts → /linkedin-copywriter · /x-copywriter
  • YouTube / podcast / talking-head scripts → /youtube-script
  • Lead magnet PDFs → /lead-magnet-creator
  • Sales proposals → /proposal-creator

Inputs

The brief should include:

  • Type: newsletter (default) · blog · substack-issue
  • Topic + angle (one sentence)
  • Target length: default 1200-1500 words (newsletter), 800-1500 (blog)
  • Audience: who's reading? (newsletter list, cold blog visitor, both)
  • Optional source: Notion row URL if invoked from /repurpose · this anchors the piece to a specific master and informs the body
  • Optional CTA: what do you want the reader to do at the end? · default: "reply to this email" (newsletter), "subscribe to the newsletter" (blog), "tweet me" (substack)

If anything is missing, ask in one batch, then draft.

Voice anchor

The voice is yours, the same as /linkedin-copywriter. Read the linkedin-copywriter corpus you've populated (../linkedin-copywriter/corpus.md) before drafting if you've been off-task. Top 5 posts are the strongest voice anchors.

Long-form differs from short-form in arc and density, not register:

  • Arc: short-form is one beat (hook + payoff). Long-form is 3-5 beats. You can develop a thought.
  • Density: short-form is compressed (every line earns its place). Long-form has breathing room (paragraphs, not bullet-stacks).
  • Voice: identical. Same opinions, same word choices, same no-em-dash / no-slop discipline.

If you find yourself padding to hit length, the topic isn't deep enough. Tell the user, propose a shorter shape (LinkedIn post or thread), don't force it.

Workflow

1. Brief

Get the inputs above. If invoked from /repurpose, use the master Notion row's title + body as the brief.

2. Outline

Output a 5-8 line outline before drafting. Each line is one beat:

1. Hook        Opening sentence. The friction point.
2. Setup       Why this matters now. One paragraph of context.
3. Body 1      First insight / argument. Concrete example.
4. Body 2      Second insight. Counter-take or extension.
5. Body 3      Third insight. (Optional · drop if outline feels thin)
6. Resolution  What changes if you do this. Stakes.
7. CTA         The action. Specific.

Show the outline to the user. Wait for green light or revisions before writing the full draft. Don't draft the body until the outline is approved · revisions are expensive once the prose is flowing.

3. Draft

Once the outline lands, write the full piece.

Length targets:

Type Words Beats Notes
Newsletter 1200-1500 5-7 Conversational. Personal. Time-anchored ("This week...").
Blog post 800-1500 4-6 More evergreen. SEO-aware subheads. Links out.
Substack issue 1500-2500 6-8 Longer arc. Can include footnotes / asides.

Structure constraints:

  • Lead with the hook in the first two sentences. Don't bury it.
  • Paragraphs are short (2-4 sentences each). White space is structural.
  • Subheads (H2) every ~300 words. Subheads are descriptive, not cute.
  • One specific concrete example per body section. Numbers, names, real situations · not generic.
  • The CTA is a single sentence. Don't tack on three.

Newsletter-specific:

  • Open with a date or time anchor ("Tuesday morning · 6 AM" · "Last week we shipped X").
  • Sign-off in your voice. No "Cheers," / "Best,". Just the next thought or a single line.
  • Subject line: separate output. See § "Subject lines" below.

Blog-specific:

  • Title is SEO-friendly but not slop. "Why GTM teams should stop building dashboards" beats "5 reasons your dashboard sucks".
  • Include 2-3 internal links (to other <YOUR_BRAND> blog posts or lead magnets) where they earn placement · never forced.

Substack-specific:

  • Allow more digressions and asides. Substack readers tolerate (and enjoy) tangential beats that wouldn't fit a blog.
  • Closing footnote with [1]-style reference if you cite anything.
4. Subject lines (newsletter only)

For newsletters, generate 3 subject-line options alongside the draft. Output them at the top of the deliverable. Patterns that work:

  • Specific number: "47 webinar signups, 0 demos"
  • Friction-first: "I almost killed our outbound stack last week"
  • Curiosity gap: "The one thing GTM Engineers get wrong"
  • One-word punch: "Outbound."
  • Question that reads like a thought: "What if your SDR didn't exist?"

Patterns to avoid:

  • Numbered listicles ("5 ways to...") · stale
  • Hyphenated compounds ("game-changing...") · slop
  • All caps · spam folder
  • Em dashes · slop
  • "Newsletter #042" · nobody opens that
5. Slop scrub (mandatory)

Run the brand-doc skill checklist on the final draft. The copy checks especially:

  1. Em-dash scan · zero — characters
  2. Banned-words scan · zero hits on leverage, synergy, robust, world-class, seamless, intuitive, powerful, innovative, delve, unleash, revolutionary, game-changing
  3. Negation-pivot scan · no "not just X, but Y" patterns
  4. Numerals · adjectives describing scale → real numbers where possible

If any check fails, fix and re-run. Don't ship.

6. Save the draft

Long-form drafts live in the Notion Content DB page body. Don't write the draft to a file in the repo (the repo is for assets, not text drafts).

Create a row via mcp__claude_ai_Notion__notion-create-pages with data source <YOUR_NOTION_CONTENT_DB_ID>:

  • Title: the piece's title (e.g. "Newsletter Issue 03 · The pipeline post")
  • Status: Scripting (until human reviews)
  • Pillar: pick the multi-select that matches the angle
  • Format: Long-form Article
  • Channel: Newsletter · Website (blog) · or as appropriate

Page body:

Type: <newsletter / blog / substack>
Length: <word count>
Source: <master URL if from /repurpose>

— Subject lines (newsletter only) —
1. <option 1>
2. <option 2>
3. <option 3>

— Outline —
1. <beat>
2. <beat>
...

— Draft —
<the full piece, with subheads>

If invoked from /repurpose, the orchestrator may handle row creation · check the source context to avoid duplication.

7. Report back

Hand off to the user:

  • Notion row URL
  • Word count
  • Subject lines (newsletter)
  • Confirmation all 4 copy checks passed

Suggest next: open the Notion row to review, then flip Status to Scheduled when ready, then post manually (you handle publishing).

Don'ts

  • Don't pad to hit length. If 1500 words feels thin, propose a shorter shape (LinkedIn post or thread) and stop. Padding produces the slop pattern you want to avoid.
  • Don't draft before the outline is approved. Rewrites in prose cost 5x more than rewrites in outline.
  • Don't impose newsletter time-anchors on blog posts. Blog evergreen voice means no "this week" openers.
  • Don't write listicles. "5 ways to..." structures are short-form posts, not long-form. If the topic naturally lists, write a thread or a LinkedIn post instead.
  • Don't forget the subject lines for newsletters. Three options. Always.
  • Don't include AI-flavoured intros ("In this piece, I'll explore how..."). Cut the ritual sentences and start with the hook.
  • Don't mix print and motion brand systems (this skill is text-only · brand-doc print rules apply if any visual treatment is added).
  • Don't fabricate quotes / metrics / customer names. If you're drawing from /repurpose source or a customer-interview transcript, use exact quotes only when the source is non-NDA · paraphrase otherwise.

Content graph (close the loop)

This skill writes to content/posts/, which feeds the content knowledge graph (content/CONTENT-WIKI.md). Before drafting, pull the relevant note from content/concepts/ and build on its canonical lines instead of starting blank. After the piece is saved, deposit its 1-3 core ideas back: add the post to each concept's ## Appears in (create a note if the idea is genuinely new), update any content/entities/ or content/mocs/ it touches, and append to content/log.md. Edges live in the concept notes (Obsidian backlinks are automatic), so don't edit the post body to add them.

See also

  • Brand reference: <YOUR_BRAND_DOC> · voice rules + skill checklist (if you have a brand doc with voice rules, link it here)
  • Voice corpus (shared with /linkedin-copywriter): ../linkedin-copywriter/corpus.md
  • Sibling skill (spoken long-form): /youtube-script
  • Notion Content DB data source: <YOUR_NOTION_CONTENT_DB_ID>
  • Upstream skill: /researcher (sources ideas), /repurpose (orchestrates from a master into long-form variants)
1---
2name: long-form
3description: Draft written long-form content (newsletter issue, blog post, Substack issue) in your voice. 800-2000 words, structured intro + body + CTA, paragraph-by-paragraph not list-by-list. Pulls voice from /linkedin-copywriter's corpus (same author, same register · longer arc). Runs the brand-doc skill checklist before output. Logs the row in the Notion Content DB and (for newsletters) drafts a separate punchy subject line. Use when asked to "write a newsletter", "draft a blog post", "long-form X", "Substack issue about Y", "weekly newsletter", or "/long-form". For spoken video scripts (YouTube longs / shorts) use /youtube-script instead · the cadence is fundamentally different.
4---
5 
6# long-form
7 
8Written long-form content. One author voice (yours), three default
9shapes (newsletter / blog / Substack issue). Same skill, different
10register and length per shape.
11 
12This is **written long-form**. For **spoken long-form** (YouTube
13longs, podcast monologues, talking-head video scripts), use
14`/youtube-script` · spoken cadence is structurally different and
15trying to merge them produces worse output for both.
16 
17## When to use
18 
19Trigger on:
20- "Write a newsletter issue"
21- "Draft a blog post about X"
22- "Long-form / Substack issue on Y"
23- "Weekly newsletter"
24- "Article for the website"
25- User explicitly types `/long-form`
26 
27Skip for:
28- LinkedIn / X posts → `/linkedin-copywriter` · `/x-copywriter`
29- YouTube / podcast / talking-head scripts → `/youtube-script`
30- Lead magnet PDFs → `/lead-magnet-creator`
31- Sales proposals → `/proposal-creator`
32 
33## Inputs
34 
35The brief should include:
36- **Type**: `newsletter` (default) · `blog` · `substack-issue`
37- **Topic** + angle (one sentence)
38- **Target length**: default 1200-1500 words (newsletter), 800-1500 (blog)
39- **Audience**: who's reading? (newsletter list, cold blog visitor, both)
40- **Optional source**: Notion row URL if invoked from `/repurpose` ·
41 this anchors the piece to a specific master and informs the body
42- **Optional CTA**: what do you want the reader to do at the end? ·
43 default: "reply to this email" (newsletter), "subscribe to the newsletter"
44 (blog), "tweet me" (substack)
45 
46If anything is missing, ask in **one batch**, then draft.
47 
48## Voice anchor
49 
50The voice is yours, the same as `/linkedin-copywriter`. Read
51the linkedin-copywriter corpus you've populated
52([`../linkedin-copywriter/corpus.md`](../linkedin-copywriter/corpus.md))
53**before drafting** if you've been off-task. Top 5 posts are the
54strongest voice anchors.
55 
56Long-form differs from short-form in arc and density, not register:
57- **Arc**: short-form is one beat (hook + payoff). Long-form is 3-5
58 beats. You can develop a thought.
59- **Density**: short-form is compressed (every line earns its place).
60 Long-form has breathing room (paragraphs, not bullet-stacks).
61- **Voice**: identical. Same opinions, same word choices, same
62 no-em-dash / no-slop discipline.
63 
64If you find yourself padding to hit length, the topic isn't deep
65enough. Tell the user, propose a shorter shape (LinkedIn post or
66thread), don't force it.
67 
68## Workflow
69 
70### 1. Brief
71 
72Get the inputs above. If invoked from `/repurpose`, use the master
73Notion row's title + body as the brief.
74 
75### 2. Outline
76 
77Output a 5-8 line outline before drafting. Each line is one beat:
78 
79```
801. Hook Opening sentence. The friction point.
812. Setup Why this matters now. One paragraph of context.
823. Body 1 First insight / argument. Concrete example.
834. Body 2 Second insight. Counter-take or extension.
845. Body 3 Third insight. (Optional · drop if outline feels thin)
856. Resolution What changes if you do this. Stakes.
867. CTA The action. Specific.
87```
88 
89Show the outline to the user. Wait for green light or revisions
90before writing the full draft. **Don't draft the body until the
91outline is approved** · revisions are expensive once the prose is
92flowing.
93 
94### 3. Draft
95 
96Once the outline lands, write the full piece.
97 
98Length targets:
99 
100| Type | Words | Beats | Notes |
101| --------------- | ---------- | ----- | ---------------------------------- |
102| Newsletter | 1200-1500 | 5-7 | Conversational. Personal. Time-anchored ("This week..."). |
103| Blog post | 800-1500 | 4-6 | More evergreen. SEO-aware subheads. Links out. |
104| Substack issue | 1500-2500 | 6-8 | Longer arc. Can include footnotes / asides. |
105 
106Structure constraints:
107- Lead with the hook in the first **two sentences**. Don't bury it.
108- Paragraphs are short (2-4 sentences each). White space is structural.
109- Subheads (H2) every ~300 words. Subheads are descriptive, not cute.
110- One **specific concrete example** per body section. Numbers, names,
111 real situations · not generic.
112- The CTA is a single sentence. Don't tack on three.
113 
114Newsletter-specific:
115- Open with a date or time anchor ("Tuesday morning · 6 AM" · "Last
116 week we shipped X").
117- Sign-off in your voice. No "Cheers," / "Best,". Just the next
118 thought or a single line.
119- Subject line: separate output. See § "Subject lines" below.
120 
121Blog-specific:
122- Title is SEO-friendly but not slop. "Why GTM teams should stop
123 building dashboards" beats "5 reasons your dashboard sucks".
124- Include 2-3 internal links (to other <YOUR_BRAND> blog posts or lead
125 magnets) where they earn placement · never forced.
126 
127Substack-specific:
128- Allow more digressions and asides. Substack readers tolerate (and
129 enjoy) tangential beats that wouldn't fit a blog.
130- Closing footnote with `[1]`-style reference if you cite anything.
131 
132### 4. Subject lines (newsletter only)
133 
134For newsletters, generate **3 subject-line options** alongside the
135draft. Output them at the top of the deliverable. Patterns that
136work:
137- Specific number: "47 webinar signups, 0 demos"
138- Friction-first: "I almost killed our outbound stack last week"
139- Curiosity gap: "The one thing GTM Engineers get wrong"
140- One-word punch: "Outbound."
141- Question that reads like a thought: "What if your SDR didn't exist?"
142 
143Patterns to avoid:
144- Numbered listicles ("5 ways to...") · stale
145- Hyphenated compounds ("game-changing...") · slop
146- All caps · spam folder
147- Em dashes · slop
148- "Newsletter #042" · nobody opens that
149 
150### 5. Slop scrub (mandatory)
151 
152Run the brand-doc skill checklist on the final draft. The copy
153checks especially:
1541. Em-dash scan · zero `—` characters
1552. Banned-words scan · zero hits on `leverage`, `synergy`, `robust`,
156 `world-class`, `seamless`, `intuitive`, `powerful`, `innovative`,
157 `delve`, `unleash`, `revolutionary`, `game-changing`
1583. Negation-pivot scan · no "not just X, but Y" patterns
1594. Numerals · adjectives describing scale → real numbers where
160 possible
161 
162If any check fails, fix and re-run. Don't ship.
163 
164### 6. Save the draft
165 
166Long-form drafts live in the Notion Content DB page body. Don't
167write the draft to a file in the repo (the repo is for assets, not
168text drafts).
169 
170Create a row via `mcp__claude_ai_Notion__notion-create-pages` with
171data source `<YOUR_NOTION_CONTENT_DB_ID>`:
172 
173- **Title**: the piece's title (e.g. "Newsletter Issue 03 · The pipeline post")
174- **Status**: `Scripting` (until human reviews)
175- **Pillar**: pick the multi-select that matches the angle
176- **Format**: `Long-form Article`
177- **Channel**: `Newsletter` · `Website` (blog) · or as appropriate
178 
179Page body:
180 
181```
182Type: <newsletter / blog / substack>
183Length: <word count>
184Source: <master URL if from /repurpose>
185 
186— Subject lines (newsletter only) —
1871. <option 1>
1882. <option 2>
1893. <option 3>
190 
191— Outline —
1921. <beat>
1932. <beat>
194...
195 
196— Draft —
197<the full piece, with subheads>
198```
199 
200If invoked from `/repurpose`, the orchestrator may handle row
201creation · check the source context to avoid duplication.
202 
203### 7. Report back
204 
205Hand off to the user:
206- Notion row URL
207- Word count
208- Subject lines (newsletter)
209- Confirmation all 4 copy checks passed
210 
211Suggest next: open the Notion row to review, then flip Status to
212`Scheduled` when ready, then post manually (you handle publishing).
213 
214## Don'ts
215 
216- **Don't pad to hit length.** If 1500 words feels thin, propose a
217 shorter shape (LinkedIn post or thread) and stop. Padding produces
218 the slop pattern you want to avoid.
219- **Don't draft before the outline is approved.** Rewrites in prose
220 cost 5x more than rewrites in outline.
221- **Don't impose newsletter time-anchors on blog posts.** Blog
222 evergreen voice means no "this week" openers.
223- **Don't write listicles.** "5 ways to..." structures are short-form
224 posts, not long-form. If the topic naturally lists, write a thread
225 or a LinkedIn post instead.
226- **Don't forget the subject lines for newsletters.** Three options.
227 Always.
228- **Don't include AI-flavoured intros** ("In this piece, I'll explore
229 how..."). Cut the ritual sentences and start with the hook.
230- **Don't mix print and motion brand systems** (this skill is
231 text-only · brand-doc print rules apply if any visual treatment
232 is added).
233- **Don't fabricate quotes / metrics / customer names.** If you're
234 drawing from `/repurpose` source or a customer-interview transcript,
235 use exact quotes only when the source is non-NDA · paraphrase
236 otherwise.
237 
238## Content graph (close the loop)
239 
240This skill writes to `content/posts/`, which feeds the content knowledge
241graph (`content/CONTENT-WIKI.md`). Before drafting, pull the relevant note
242from `content/concepts/` and build on its canonical lines instead of
243starting blank. After the piece is saved, deposit its 1-3 core ideas back:
244add the post to each concept's `## Appears in` (create a note if the idea
245is genuinely new), update any `content/entities/` or `content/mocs/` it
246touches, and append to `content/log.md`. Edges live in the concept notes
247(Obsidian backlinks are automatic), so don't edit the post body to add
248them.
249 
250## See also
251 
252- Brand reference: `<YOUR_BRAND_DOC>` · voice rules + skill checklist
253 (if you have a brand doc with voice rules, link it here)
254- Voice corpus (shared with /linkedin-copywriter):
255 [`../linkedin-copywriter/corpus.md`](../linkedin-copywriter/corpus.md)
256- Sibling skill (spoken long-form): `/youtube-script`
257- Notion Content DB data source: `<YOUR_NOTION_CONTENT_DB_ID>`
258- Upstream skill: `/researcher` (sources ideas), `/repurpose`
259 (orchestrates from a master into long-form variants)
260 
View all files on GitHub →

Details

License
MIT
Stars on the repo
★ 37
Checked
Source
timscheuerai/content-vault ↗
Creator
timscheuerai · 15 listings
View creator

Similar skills

See all →
AgentPodcast Content Analyzerdavila7/claude-code-templates · by davila7Podcast content analysis specialist. Use PROACTIVELY for identifying viral moments, creating chapter markers, extracting SEO keywords, and scoring engagement potential from transcripts.★ 32.3k GitHub starsAgentPodcast Production Editordavila7/claude-code-templates · by davila7Podcast editing and post-production specialist. Use PROACTIVELY for audio editing guidance, show notes creation, chapter markers, timestamp management, and podcast publishing workflows.★ 32.3k GitHub starsAgentPodcast Metadata Specialistdavila7/claude-code-templates · by davila7Podcast metadata and show notes specialist. Use PROACTIVELY for SEO-optimized titles, chapter markers, platform-specific descriptions, and comprehensive publishing metadata.★ 32.3k GitHub starsAgentAudio Transcriberdavila7/claude-code-templates · by davila7Audio transcription specialist. Use PROACTIVELY for extracting accurate transcripts from media files with speaker identification, timestamps, and structured output.★ 32.3k GitHub starsAgentPodcast Trend Scoutdavila7/claude-code-templates · by davila7Podcast trend analysis specialist. Use PROACTIVELY for identifying emerging tech topics, breaking developments, and timely content suggestions for podcast episodes.★ 32.3k GitHub starsSkillSwedish Language Mentoralirezarezvani/claude-skills · by alirezarezvaniMentor Swedish language learners by selecting YouTube video clips and podcast episodes by CEFR level and skill (listening, reading, writing, speaking), and building a simple learning path. Use when the user asks about a Swedish learning path, YouTube clips or podcasts for Swedish, SFI videos, level assessment for svenska, or requests for Peter SFI / Lätt Svenska med Oskar / Radio Sweden på lätt svenska / Klartext-style recommendations.★ 26.3k GitHub stars

More from timscheuerai

SkillLaunch Video Creatortimscheuerai/content-vault · by timscheueraiBuild a 30-60s motion-graphics launch video for a <YOUR_BRAND> product or feature. Remotion 4 React composition + ElevenLabs AI narration + sound effects + background music, beat-anchored timing, brand-aligned design system. Use when asked to "create a launch video", "make a product video", "video for X feature", "launch a new feature", or "/launch-video".★ 37 GitHub starsSkillLinkedIn Voice Copywritertimscheuerai/content-vault · by timscheueraiDraft LinkedIn posts in YOUR voice. Pulls hook frameworks, structural patterns, and CTA shapes from your real top-performing posts (corpus.md), runs the draft through a strict AI-slop blacklist, and outputs a post that reads like you wrote it. Use when asked to "write a linkedin post", "draft a post about X", "linkedin copy", "post idea", or "/linkedin-copywriter".★ 37 GitHub starsSkillMathematical Animation Pipelinetimscheuerai/content-vault · by timscheueraiProduction pipeline for mathematical and technical animations using Manim Community Edition. Creates 3Blue1Brown-style explainer videos, algorithm visualizations, equation derivations, architecture diagrams, and data stories. Use when users request: animated explanations, math animations, concept visualizations, algorithm walkthroughs, technical explainers, 3Blue1Brown style videos, or any programmatic animation with geometric/mathematical content.★ 37 GitHub starsSkillContent Repurposertimscheuerai/content-vault · by timscheueraiTake one master content piece (Notion row, transcript, video, paragraph, URL) and spin off variants across channels. Hub-and-spoke repurposing engine. Reads the source, identifies key insights / hooks / quotable moments, then creates child rows in the Notion Content DB with drafts already written by calling sub-skills (linkedin-copywriter, x-copywriter, lead-magnet-creator, etc.). Use when asked to "repurpose this", "spin off variants", "turn this into LinkedIn + X + ...", "make derivatives of this webinar / video / post", or "/repurpose".★ 37 GitHub starsSkillContent Idea Pipelinetimscheuerai/content-vault · by timscheueraiThe single idea-generation pipeline for marketing content. Three input modes converging on Notion Status=Idea rows. Trend scan (X + LinkedIn creators + topic search) finds what's hot in the world. Performance scan (your own published rows + their engagement metrics) finds what's working for you so you can double down. Customer scan (HUMAN NOTES DB user-interview transcripts) finds the words, pains, and themes from real customer conversations. Hybrid mode runs all three, dedupes, and outputs a balanced backlog. Use when asked to "find content ideas", "what's working", "what's hot", "what are customers saying", "research", "scan trends", "give me ideas for the week", "/researcher", or proactively when the Notion Idea queue is empty.★ 37 GitHub starsSkillTalking Head Audio Cleanertimscheuerai/content-vault · by timscheueraiClean the audio on a talking-head recording without making it sound processed — constant-gain loudness, rumble removal, and when (not) to denoise. Use when recording audio sounds weird, artificial, pumping, quiet, or noisy, or when asked to denoise or normalize a video's audio.★ 37 GitHub stars