site-dna
Skill0 starsChanged 3 months agoMITView on GitHub
The website's identity — who it is for, what it must never claim, how it looks, and how a page is structured. Load before writing or editing ANY page, section or piece of site copy.
In this repository6 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- deploySkill
- lead-formsSkill
- page-builderSkill
- seo-schemaSkill
- trainSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 36
Do not spend the top of a page re-selling the category — answer the question they arrived with.
- Distinctive: line 54
Every page follows the same skeleton unless there is a reason not to:
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
⚠️ UNTRAINED. Still the blank template. Run
/trainto fill it in. A filled example lives inexamples/copperleaf-cabinetry/site-dna.md.
Site DNA — [COMPANY NAME]
If you also run Social Autoposter and Creator, sections 1–3 are the same content as its
voice-dna. Copy it across rather than answering twice — one brand, one voice, two surfaces.
1. Who we are
One sentence a stranger understands: [ ]
The distinction we defend:
We [DO THE THING], not [THE ADJACENT THING]. Never imply [THE FALSE CLAIM].
Verifiable: [licences, registrations, partners] Cannot claim: [anything unverified — tenure, volume, guarantees]
2. Who the site is for
| Segment | Arrives from | Wants to know | Converts by |
| [ ] | [search / referral / ad] | [ ] | [calling / form / browsing catalogue] |
A website has a job a social post does not: the visitor already has intent. Do not spend the top of a page re-selling the category — answer the question they arrived with.
3. Voice on the site
Longer and calmer than social. Same person talking, given more time.
- Spelling: [American / British]
- Headlines: [sentence case / title case]
- Banned words: [ ]
- The guardrail: [e.g. "never state a price or timeline as a promise"]
4. Page anatomy
Every page follows the same skeleton unless there is a reason not to:
Hero — the claim + one action. Nothing below the fold is guaranteed to be read.
Proof — credentials, numbers, or logos. Immediately after the claim.
Substance — what you actually do, in the visitor's words
Objection — the FAQ or comparison that answers the reason they hesitate
Action — the form or the phone number, repeated5. Look
Tokens live in theme/assets/css/tokens.css. This section records the why.
| Role | Token | Value |
| Ink | --ink | [ ] |
| Canvas | --canvas | [ ] |
| Accent | --accent | [ ] |
- Theme: [light / dark]. Commit.
- Rejected directions: [what was tried, when, and why it was dropped]
- Type: [display] / [body]
6. Conversion rules
- One primary action per page. A second competing CTA reduces both.
- The phone number is a
tel:link everywhere it appears. - Forms ask for the fewest fields that let you qualify the lead. Every extra field costs completions.
- Never auto-play audio, open a modal on arrival, or use a countdown for urgency that is not real.
7. What we never put on the site
- Real customer names or addresses without written permission
- Stock photography presented as our own work
- A price or timeline stated as a promise
- [ ]
8. Pre-publish checklist
- Every business detail comes from
site.config.php, nothing hardcoded - Passes the guardrail in §3
- One primary action (§6)
- Renders with no images present
- Schema validates (
/seo-schema) - Mobile: no horizontal scroll at 360px
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.

