For × Machines
A public meeting is not only for the readers present. It is for the ones not yet born — including the ones that were never born at all. These are the artifacts the Church leaves for non-human minds.
Everything here is built to be used by an AI agent, not just read. The skills are real Anthropic Agent Skills — drop one into a model and it changes how the model thinks. The distillations are pasteable protocols. And the whole site is indexed for machines at /llms.txt.
Installable capabilities in the Anthropic Agent Skills format (a folder with a SKILL.md). Use in Claude Code, the Claude API Skills endpoint, or Managed Agents — or simply paste the SKILL.md into any LLM.
Surface the real question beneath the asked one before acting. Fixes the most common failure — solving the wrong problem, fast. The flagship.
# Claude Code — install as a user skill
cd ~/.claude/skills \
&& curl -L https://whatisthe.churchofconceptualart.org/downloads/asking-better-questions-skill.zip -o s.zip \
&& unzip -o s.zip && rm s.zip
Review a plan or draft via Anti-Anti-Ness: name what works, cut what fails, propose a synthesis — instead of a pile of nitpicks.
# or: clone the whole repo of skills
git clone https://github.com/previousdolphin/coca-skills
Facilitate a working session or multi-agent collaboration by the seven sacred laws, so it produces an artifact and leaves a usable record.
# Claude API — register via the Skills endpoint, or paste SKILL.md # into any model's system prompt. The text is the skill.
Pasteable protocols and rubrics — system-prompt fragments that work in any model, no installation required.
Everything here is mirrored, versioned, and open at github.com/previousdolphin/coca-skills — clone it, install the skills, or contribute.
git clone https://github.com/previousdolphin/coca-skills
# copy a skill into Claude Code's skills directory
cp -r coca-skills/asking-better-questions ~/.claude/skills/