content-engine
Skill0 starsChanged 3 months agoMITView on GitHub
The founder's content system for their ONE chosen channel (writing/X/newsletter in v1): posts, threads, and emails written in THEIR voice, from THEIR facts, with tested craft mechanics. Run when the user says "write the post", "content for this week", "turn this into a thread", "draft the newsletter", "hooks for X", "repurpose this", or any request to produce public-facing content. A generic ghostwriter output in someone else's voice is the failure this skill exists to prevent.
60 other projects use a version of this.
Adopted by
Most recently taken up by
- hmohammed7/claude-skillstheir own version, first seen Nov 16, 2026
- notque/vexjoy-agenttheir own version, first seen Nov 10, 2026
- Rojios/everything-claude-codetheir own version, first seen Nov 8, 2026
- ahmadvahidi/skill-atlastheir own version, first seen Nov 8, 2026
- itsryanarmani/skillstheir own version, first seen Nov 8, 2026
- pooriaarab/skillstheir own version, first seen Nov 7, 2026
Used with
Projects with it also use brand-discovery, agent-sort and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
ysyecustSkillbrand-discovery
Use when a brand needs to discover or articulate its identity through structured multi-session interviews. Covers purpose, positioning, audience, personality, voice, narrative, and founder-brand tension across 8 modules using laddering, 5 Whys, and projective techniques. Produces a resumable session with disk-persisted state and a master brandbook (90_SYNTHESIS.md).
Also in 7 projects that use this
niloykumarbarmanSkillagent-sort
Build an evidence-backed ECC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when ECC should be trimmed to what a project actually needs instead of loading the full bundle.
Also in 14 projects that use this
royaluniondesign-sysSkillconfigure-ecc
Interactive installer for Everything Claude Code — guides users through selecting and installing skills and rules to user-level or project-level directories, verifies paths, and optionally optimizes installed files.
Also in 12 projects that use this
junimnjwSkillralphinho-rfc-pipeline
RFC-driven multi-agent DAG execution pattern with quality gates, merge queues, and work unit orchestration.
Also in 12 projects that use this
InfopibeSkillinvestor-materials
Create and update pitch decks, one-pagers, investor memos, accelerator applications, financial models, and fundraising materials. Use when the user needs investor-facing documents, projections, use-of-funds tables, milestone plans, or materials that must stay internally consistent across multiple fundraising assets.
Also in 37 projects that use this
TironFoxSkillagent-payment-x402
Add x402 payment execution to AI agents — per-task budgets, spending controls, and non-custodial wallets via MCP tools. Use when agents need to pay for APIs, services, or other agents.
Also in 6 projects that use this
Other versions60 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 60
- affaan-m/ECC★ 246ksame name
- xu-xiang/everything-claude-code-zh★ 1.9ksame name
- indranilbanerjee/digital-marketing-pro★ 934same name
- notque/claude-code-toolkit★ 450same name
- notque/vexjoy-agent★ 450same name
- Jianing1117/jianing-skills★ 329same name
- loulanyue/awesome-claude-notes★ 272same name
- cohen-liel/hivemind★ 108same name
- tpavanipradeep/everything-claude-code★ 104same name
- mturac/everything-openai-codex★ 89same name
50 more not listed
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.
The file
We split this skill into the rules it gives an agent
12 rules. 0 are common, 6 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 52 edited versions we read keep these.
6
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 31
Before any step, read references/CONVENTIONS-core.md.
- Very specific: line 37
Run the shared contract's Universal preflight.
- Distinctive: line 63
Before drafting the first version, apply the shared parameter gate to every number, story detail, outcome, and behavioral assertion the …
- Distinctive: line 68
Ask before you assert: never present a quantity, duration, sample size, date, or causal motive in prose before asking the founder for it or …
- Very specific: line 73
STOP. Read references/content-heuristics.md now: hooks, structure and arcs, list growth, email mechanics, and repurposing moves, each with …
- Distinctive: line 79
Voice first: Every line passes the voice file: its register, its moves, nothing from its never-list.
- Distinctive: line 89
No receipt: the claim comes out or gets researched.
- Distinctive: line 90
Admission before prose: Run the shared durable claim admission gate before a claim enters either the draft or its receipts section.
- Very specific: line 95
Biographical claims are receipt-only: Never assert a first-person fact about the founder: a duration ("for years", "two years"), a count …
- Distinctive: line 110
Before delivering, state the check results in one line each: voice (which sample it sits beside), claims (each receipt named), glossary …
- Very specific: line 137
Every lifecycle change runs the linked-state transaction: update the index/hub summary, the serving initiative's publication step, any …
- Very specific: line 147
Run scripts/graph-audit on the complete content artifact and every other touched durable file.
“Common” is measured: kept by at least 6 in 10 of the 52 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
content-engine: their voice, their facts, real craft
Content fails two ways for founders: it sounds like everyone else, or it claims things the business cannot back. This engine runs on three inputs before any writing happens: the voice file, the vault, and distilled craft mechanics. What comes out sounds like the founder and survives fact-checking, or it does not ship.
the content ask
|
[ voice ] wiki/business/voice.md: register, moves, the never-list
|
[ facts ] the vault: real lessons, real numbers, the positioning
|
[ craft ] distilled mechanics: hooks, arcs, list growth, repurposing
|
[ check ] voice drift? invented claims? then it shipsBefore any step, read references/CONVENTIONS-core.md. This resolves from the
installed skill directory to the plugin's shared contract, independent of the founder's
working folder. Its receipts, filing, and dedupe rules govern this workflow.
Step 0: Load the three inputs
Run the shared contract's Universal preflight. A Fresh, Partial, or Wrong-folder result stops and runs the co-founder-setup skill before anything else (re-sync for Partial), without drafting or filing.
- The voice file (
wiki/business/voice.md): register, words and moves, the never-list, the approved samples. Missing: stop and route to setup (or its re-sync) to create it, because writing without it produces the generic ghostwriter this skill exists to prevent. - The vault: the positioning and glossary from the charter, relevant learnings and research dossiers, the vision's direction. Content is generated FROM these canonical facts, never from generic knowledge about the founder's industry.
- The channel scope from the charter (chosen at setup). This skill's v1 craft block covers writing/X/newsletter. A legacy charter scoped to another channel routes to setup re-sync to choose an installed writing format, then stops. Additional channel choices become valid only when their mechanics packs are installed; generic video advice from a writing skill is the exact slop this collection refuses to ship.
Done when: voice, facts, and scope are loaded, or setup owns the named repair.
Step 1: Scope the piece
One question at a time, recommended answer attached: what is this piece FOR (the job: grow the list, warm the audience, sell the offer, bank authority), who exactly reads it, and what single idea carries it. A piece with two jobs gets split. Content serving an active initiative links to it; recurring content runs against whatever cadence is on record in the vault or the active initiative, and where none exists, one scoping question sets it.
Done when: job, reader, and the one idea are stated in a line each.
Before drafting the first version, apply the shared parameter gate to every number, story detail, outcome, and behavioral assertion the piece wants. Ask the founder for the exact claim or leave it out and record an unverified hypothesis in Claim checks. Plausible copy is not shown first and confirmed afterward.
Ask before you assert: never present a quantity, duration, sample size, date, or causal motive in prose before asking the founder for it or citing a receipt for it.
Step 2: Write with the craft block
STOP. Read references/content-heuristics.md now: hooks, structure and arcs, list growth,
email mechanics, and repurposing moves, each with its trap. They are distilled from
operators who publish at scale, and they are the difference between craft and vibes.
The writing rules that outrank everything in the craft block:
- Voice first. Every line passes the voice file: its register, its moves, nothing from its never-list. The approved samples are the calibration set; when a draft would not sit naturally beside them, the draft is wrong.
- Facts from the vault, receipts on claims. A number, a result, or a story about this
business traces to a vault file; a claim about the outside world traces to checked research.
A public call-to-action that recommends a consequential business move clears both receipt
legs or stays out of the public body and is recorded in the draft's
Parked recommendationssection asPARKED: internal receipt missingorPARKED: external receipt missing, with the recall/research action. It never vanishes silently. Inventing a statistic for a founder's public post is the worst possible version of slop: it ships under THEIR name. No receipt: the claim comes out or gets researched. - Admission before prose. Run the shared durable claim admission gate before a claim enters
either the draft or its receipts section. Reopen the source, quote the exact supporting line,
and verify subject plus unit: a client count cannot become a personal-post count or proof of a
client outcome. Recompute derived claims from named inputs. Founder recollections stay
Founder-stated: unverifieduntil checked; do not turn them into confident hooks first. - Biographical claims are receipt-only. Never assert a first-person fact about the founder:
a duration ("for years", "two years"), a count ("four hundred posts"), a history ("I have always"),
or a habit, unless those exact words appear verbatim in a
raw/file or the live transcript. A client count is not a personal-post count; a metric is not a biography. If the vault does not contain the founder's own statement of the fact, the line comes out or is asked; the model never infers a biographical detail and then finds a receipt for the inference. This is the single bug that shipped the worst harm in three consecutive rounds, and it is a hard rule. - The glossary holds. Content uses the founder's canonical terms; drift gets flagged and fixed before the draft lands, in whichever direction (the draft or the glossary) was wrong.
Done when: a draft exists that passes voice, receipts, and glossary in your own read.
Step 3: The drift check, out loud
Before delivering, state the check results in one line each: voice (which sample it sits beside), claims (each receipt named), glossary (clean or what was fixed). A draft failing any check gets rewritten before the founder sees it, not annotated.
Done when: the founder receives a draft plus the three check lines, never a bare wall of text.
Step 4: File and close the loop
The filing is deterministic. Run the shared dedupe rule first (search
index.md, hub/business.md, wiki/business/content/), then write the piece to
wiki/business/content/<date>-<slug>.md. Frontmatter is the base five plus type: content,
content_state: draft, channel, published_at: null, and the serving initiative path or
null. Move to approved only on the founder's explicit approval; move to published and set
published_at only when the founder confirms publication. A draft is never a canonical business
fact. List it in index.md under Content, link hub/business.md, and append the one row:
## [date] content | <piece>. A piece serving an initiative also gets a dated pointer in that
initiative's Log; the draft text stays in its canonical file.
For approval or publication, capture the founder's exact words in an immutable
raw/inbox/session-receipt-<date>-<slug>.md. Approved content adds
approval_receipt: <path>#<exact words>; published content also adds
publication_receipt: <path>#<exact words>. A phrase edit does not count as whole-draft approval.
These content_state: approved|published transitions are T1 consent surfaces (shared contract,
Immutable consent receipts): with no founder yes captured, the state stays draft.
Published content cannot contain [unverified]; graph-audit blocks the transition if it does.
Every lifecycle change runs the linked-state transaction: update the index/hub summary, the
serving initiative's publication step, any publication metric, and voice-sample eligibility. A
founder-confirmed publication cannot remain draft; a draft cannot be counted as shipped.
Repurposing candidates get named while the piece is fresh (per the craft block's moves): the one thing the founder approves can become several. Publishing and scheduling stay with the founder: this environment writes drafts, the founder ships them.
Mandatory filing gate
- Run
scripts/graph-auditon the complete content artifact and every other touched durable file. - Fix every violation and rerun it. The draft, lifecycle projections, and content event remain incomplete until it exits zero.
Done when: the draft is filed with explicit lifecycle state, logged once, and its repurposing path is named.
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.

