Anthropic Official Claude Skills

Anthropic Just Published Official Claude Skills

🎬 Signal Coders 📅 Aug 12, 2026 ⏱ 20:20
Anthropic Claude Skills open source skill craft production

📦 What Actually Shipped — 17 Skills, Two Families

For a month, every big name published skill repositories — except the company that invented the format. This week Anthropic finally shipped. The repository contains 18 skill files: 17 real skills plus a template so small it can be read in full. They split into two families. 2:21

Family One — The Production Document Skills (4)

Word documents, PDFs, slide decks, spreadsheets. The README states something Signal Coders has "never seen a lab do": "These are the skills that power Claude's document capabilities under the hood," shared "as a reference for more complex skills that are actively used in a production AI application." 2:33

Sit with that. When Claude builds you a slide deck, these are the exact instructions it follows. "Every judgment call about how an agent should produce a professional document, written by the people with the most skin in the game, tested by more real-world runs than any instruction file in history — readable this morning by you." 2:58

Family Two — The 13 Example Skills (Builders)

A skill for creating skills, one for building tool servers, web artifacts, testing web apps, creative systems, canvas design, algorithmic art themes — even an animated emoji maker. Plus internal communications, brand guidelines, and document co-authoring. 3:11

🔓 The Licence Twist — Who Pays

The two families ship under different licences. The 13 examples are Apache 2.0 — genuinely open. But the four production document skills carry a different front matter: "License: proprietary.txt — has complete terms." The README states it plainly: "Source available, not open source." 3:32

The invoice, printed in the front matter: "You may read the crown jewels. You may learn from them. Running off and reselling them is a different conversation." The 13 Apache skills grow the format (the format's author wins when the format wins). The 4 production skills are protected by licence. And the README's quiet line: "These example skills are already available to paid plans in claude.ai." Generosity with the knowledge, a licence on the production assets, the convenience sold as a plan. "Nothing hidden. It's all in the files." 15:16

🏗️ The Architecture — Three Layers, Published Numbers

The best file in the repo is the skill that teaches skill-making, because it writes down "the actual context arithmetic everyone else handwaves." The three-level loading system, verbatim: 4:02

LevelContentCost
1 — MetadataName + description, always in context~100 words — "the permanent tax"
2 — BodyLoaded only when the skill triggersUnder 500 lines ideal
3 — Bundled resourcesScripts, loaded as neededUnlimited — "scripts can execute without loading"
Context is a budget, and the entire official design spends it late. "Across the 17 real skills, the median is 129 lines. Half the official set is shorter than a college essay." Community files routinely run several hundred lines before their first concrete instruction. "A script can do its work without ever entering the model's context. A thousand lines of battle-tested Python costs zero words of attention." 5:00

The whole spec distilled: the template is six lines — name, description ("describe the skill and when Claude should use it"), a heading that says "Insert instructions below." That's the winning format of the format wars. 5:24

🎯 The Description Doctrine

If the official set has one obsession, it's the description field. The skill-creator file calls it "the primary triggering mechanism": "Include both what the skill does and specific context for when to use it. All when-to-use info goes here, not in the body." 5:39

They practice what they preach. The slide-deck skill's actual description: "Use this skill anytime a PPTX file is involved in any way, as input, output, or both. Trigger whenever the user mentions deck, slides, presentation. If a PPTX file needs to be opened, created, or touched, use this skill." "That's not documentation prose. That's a tripwire, engineered." 6:35

The law to cut into stone: "The description is the product. The body is the manual. A skill that never fires is indistinguishable from a skill that doesn't exist. And the home team just told you in writing that firing it is the hard part." 6:56

😅 The Confession — "Undertrigger" and "Pushy"

The sentence Signal Coders flagged as "the single most useful sentence I've read about making skills actually fire," quoted verbatim from the official file: 5:58

"Currently, Claude has a tendency to undertrigger skills — to not use them when they'd be useful. To combat this, please make the skill descriptions a little bit pushy."

"Anthropic, in an official repository, documenting a behavioral weakness of their own model and prescribing the fix. Not 'our model intelligently selects the right skill.' The honest version: it hesitates, so write descriptions that lean on it." The technique: don't write "how to build a dashboard" — write "use this whenever the user mentions dashboards, data visualization, metrics, or wants to display any kind of data, even if they don't explicitly ask for a dashboard." 6:10

⚙️ Code Over Tokens — The Move the Community Isn't Making

The deepest difference: 70 Python files in this repository. The community packs ship almost none. Open the slide-deck skill and it "barely explains anything — it's a routing table": write a script against the pre-installed library, run the converter, thumbnail.py, duplicating slides, cleaning orphaned media, validating against the format's schema with "an auto-repair flag for common breakage." 7:10

The instructions don't teach the model delicate XML surgery in prose — "they hand it instruments." A script is deterministic: "it cannot hallucinate a malformed relationship file. It does the same thing on run one and run one million, which is what you need when the output ships inside a product." 7:50

The favourite line in the entire repository — after noting the library is pre-installed, the wonderfully specific instruction: "Do not run npm install first." The file then says, verbatim: "The model knows the API. These are the foot guns." Eight words containing a whole philosophy of instruction. "Don't spend your context teaching what the model already knows — it read the documentation during training. Spend it only on the places where the model predictably breaks." 8:06
The law: teach the gaps, not the API. "The option object that mutates in place. The call that hangs in a sandbox. The text that fragments across XML runs so a phrase you can see doesn't exist as a string. The foot-gun list is the one thing training data can't give the model — because foot guns are learned by bleeding. And if it must be exact, make it a script, not a sentence." 8:45

🤫 The Strangely Quiet Guardrails

The finding that genuinely surprised Signal Coders — a number. The community files "shout capital letters everywhere: NEVER do this, ALWAYS do that, WARNING stacked like sandbags." So he ran a count across all 17 official skills: 10:34

ImperativeCount (across 17 skills)
"NEVER" (capitals)3
"ALWAYS"5
"DO NOT"1
Total raised voices9

"17 skills — including the four that run in production — and the home team raises its voice nine times across the entire set. Some community files exceed that in their first screen." How? "Constraint doesn't live in volume here. It lives in structure." The workflow is a table with one recommended tool per task, so the wrong path isn't offered. Dangerous operations are wrapped in validators that name their own fixes. Exactness lives in scripts. 11:02

"Capitals are what you reach for when you haven't built the guardrail. Mature instructions replace volume with structure. The loudest files in this gold rush are, by that measure, the least engineered." 11:28

🔍 Two Honest Violations

Score the 17 against their own rules (descriptions carry triggers, bodies under 500 lines, structure over shouting) and 15 pass clean. Two don't — and both violations are more interesting than compliance. 11:44

Violation one: the API skill is 548 lines — the longest file, and the only one over the team's own 500-line ideal. "The spec's author exceeding the spec, lightly." 12:01

Violation two: the internal-communication skill — 32 lines, the shortest real skill — is written in the first person: "A set of resources to help me write all kinds of internal communications using the formats that my company likes to use." "Help me. My company." "That is not a document written for publication. That's somebody's actual working file from inside Anthropic — published with the fingerprints still on it." 12:18

"The violations are the best evidence in the repository. Polished marketing sets don't break their own rules — they're written by committees to be admired. Working files break rules because they're used. The imperfections are the authenticity mark. This is the real toolbox tipped out onto the table, dust included." 12:36

🕳️ What's Missing — and What the Absence Reveals

Read the inventory and notice what is not there: no coding-style skills, no test-driven-development skill, no code-review skill, no debugging methodology, no personas, no "act as a senior architect," no domain packs for finance or marketing or law. "The official set is capabilities and builders." 13:10

The contrast with the month's other shelves is the finding: personal collections are essays, workflow philosophy, methodology, taste. The search giant ships job descriptions by the dozen. The home team ships short routers, 70 scripts, measured triggering, nine raised voices. 13:28

"The platform ships capabilities, the community ships opinions. Anthropic will teach the model to produce a valid spreadsheet — that's ground truth, one right answer, their responsibility. Whether your agent should write tests before code, how it should review a pull request, what personality it should wear — that's contested territory, and the home team is conspicuously declining to legislate it. The gold-rush repositories aren't competing with this set. They're the layer above it, and this release quietly ratifies their role." 13:49

Which settles where the format war stands: "The moat they're choosing isn't secrecy. It's being visibly better at writing instructions than anyone else — in public, where it doubles as the reference implementation." 14:28

✅ Five Lessons — Portable to Any Agent

  1. Write the trigger, not the essay. All the "when to use" lives in the description, and the description leans — because the author of the model told you in writing that firing is the failure mode.
  2. Spend context late. 100 words always, 500 lines on trigger, unlimited tooling that never loads. Three levels, budget accordingly.
  3. If it must be exact, make it a script. Language for judgment, code for precision. A script can't hallucinate a malformed file.
  4. Document the foot guns, not the API. The model read the docs during training. It didn't read the postmortems. "The model knows the API — these are the foot guns."
  5. Whisper. Nine raised voices across 17 production-grade files. Structure beats capitals. "Capitals are what you reach for when you haven't built the guardrail."

🔗 Resources & Links

📍 Timestamp Index

0:00 One voice stayed silent — the company that invented the format
2:21 What shipped: 17 real skills + 6-line template, two families
2:33 The production document skills — the real files running now
3:32 The licence twist: Apache vs proprietary — "source available"
4:02 A skill is three layers — 100 words / 500 lines / unlimited scripts
5:39 The description doctrine — "the description is the product"
5:58 The confession: Claude "undertriggers" — make descriptions pushy
7:10 Code over tokens: 70 scripts, the slide deck is a routing table
8:06 "The model knows the API. These are the foot guns."
10:34 The quiet guardrails — 9 raised voices across 17 skills
11:44 Two honest violations — 548-line API skill, first-person comms skill
13:10 What's missing: platform ships capabilities, community ships opinions
15:16 Who pays — the invoice printed in the front matter
18:42 Five lessons portable to any agent you run tonight
☰ View all