thought-leadership
Skill0 starsChanged 33 days agoMITView on GitHub
Writes long-form thought leadership (articles, essays, whitepapers, guest posts) built on one contrarian, specific
4 other projects use a version of this.
Adopted by
Most recently taken up by
- anotb/management-consulting-plugintheir own version, first seen Nov 1, 2026
- brainbytes-dev/everything-claude-marketingtheir own version, first seen Nov 1, 2026
- Postman-Devrel/devrel-claude-code-skillstheir own version, first seen Jul 28, 2026
- az9713/ai-co-writing-claude-skillstheir own version, first seen Dec 12, 2025
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- anotb/management-consulting-plugin★ 62same name
- az9713/ai-co-writing-claude-skills★ 61same name
- brainbytes-dev/everything-claude-marketing★ 9same name
- Postman-Devrel/devrel-claude-code-skills★ 4same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository4 other files it gives its agents.
- content-auditSkill
- content-opsSkill
- content-strategySkill
- newsletterSkill
The file
We split this skill into the rules it gives an agent
9 rules. 0 are common, 4 are distinctive, and 5 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 13
Write a long-form piece that changes what the reader believes.
- Distinctive: line 17
Use it for articles, essays, whitepapers and guest posts where there's a real point of view to argue.
- Very specific: line 19
Don't use it for social posts, newsletter copy (use /newsletter), website copy, or a topic where the user has no opinion yet.
- Very specific: line 23
Read these if they exist. Agency users: swap context/ for clients/<client>/context/.
- Distinctive: line 32
Ask the user for: the topic, the audience, the core argument if they have one, any source material (interviews, data, notes), the …
- Very specific: line 34
If context/expert-pov/ is missing, interview the user for five minutes: what do they believe that most people in their field don't, and …
- Very specific: line 59
Evidence check: no invented customers, numbers or quotes.
- Distinctive: line 68
Every claim is sourced or visibly flagged.
- Very specific: line 82
Write to outputs/content/MMYY-<topic-slug>.md: final piece, three title options, source list, open flags.
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Thought leadership
Write a long-form piece that changes what the reader believes. The method: find one governing thought the reader doesn't hold yet, build the argument for it, then write and polish, with the user approving each stage. Every claim is sourced or flagged.
When to use it
Use it for articles, essays, whitepapers and guest posts where there's a real point of view to argue.
Don't use it for social posts, newsletter copy (use /newsletter), website copy, or a topic where the user has no opinion yet. If there's no angle, help find one first or stop.
Inputs
Read these if they exist. Agency users: swap context/ for clients/<client>/context/.
context/expert-pov/: the founder's or experts' opinions, stories and arguments. The best source of angles.context/icp/: who's reading, what they believe now and what they're stuck on.context/positioning/andcontext/messaging/: what the company wants to be known for, and proof points.context/competitors/: what rivals already say, so the piece doesn't repeat it.context/brand/: voice rules.outputs/content/*-content-strategy.md: which cluster this piece belongs to.
Ask the user for: the topic, the audience, the core argument if they have one, any source material (interviews, data, notes), the publication and the target length (LinkedIn article 800-1,500 words, blog 1,500-2,500, guest post 2,000-3,500).
If context/expert-pov/ is missing, interview the user for five minutes: what do they believe that most people in their field don't, and what happened that made them believe it?
Stage 1: the argument (then stop)
- Find the governing thought. Ask: "What's the one thing you want the reader to believe after this that they didn't believe before?" It has to be contrarian (a smart reader could disagree) and specific (it names who, what and why). "AI will change marketing" fails. "Most B2B teams should cut their blog output in half and spend the time on sales calls" passes.
- Find the reader's starting point. What they believe now, the gap to the governing thought, and the three objections they'll raise.
- Build the pyramid. The governing thought on top, two to four key arguments under it, evidence under each argument. This is the Pyramid Principle (Barbara Minto).
- Test it. The key arguments don't overlap and together fully support the governing thought. Ask "so what?" of every level.
- Source check. List every claim and its source. Anything without one gets
[source needed]or comes out.
Show the outline, objections and source list. Wait for the user to approve before writing prose.
Stage 2: the draft (then stop)
- Opening: a hook that names a tension the reader feels, short context, then the governing thought stated plainly. No slow throat-clearing.
- Sections: one per key argument. Each opens with its claim, gives the evidence, answers the relevant objection, then hands off to the next.
- Closing: restate the governing thought in new words, say what the reader should do differently, and end with one call to action.
- Titles: offer three options.
Show the full draft. Wait for approval before polishing.
Stage 3: the polish (then stop)
- Voice: apply
context/brand/. Plain words, contractions, sentence-case headings, specific over superlative. - Reader check: would the ICP care, is it pitched at their level, would they forward it to a colleague?
- Evidence check: no invented customers, numbers or quotes. Unverified examples are marked
[example, verify before publishing]. External facts carry a link and the date accessed. - Read it aloud. Cut anything you wouldn't say to a peer.
Deliver the final piece with a short report: sources used, flags still open, and anything that needs the user's sign-off.
What good looks like
- The governing thought fits in one sentence and someone could argue against it.
- The key arguments are distinct and each is backed by evidence the reader can check.
- Every claim is sourced or visibly flagged. Nothing invented.
- One or two persuasion moves that are true for this reader (a specific example, a named result, a credible source), not a pile of them.
- Concrete details over abstraction: a real situation, a number, a named step.
Mistakes to avoid
- A safe thesis. If nobody would disagree, it isn't thought leadership.
- Skipping the stops. Writing prose on a weak argument wastes the user's review time.
- "Studies show" without the study.
- Buzzwords and filler: "game-changer", "cutting-edge", "in today's fast-paced world".
- Ending on a summary instead of what the reader should do next.
Output
Write to outputs/content/MMYY-<topic-slug>.md: final piece, three title options, source list, open flags.
Next step: turn the piece into a newsletter issue with /newsletter, or slot it into the calendar from /content-ops.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

