docs: progressive disclosure README — Level 0/1/2 paths#2
Conversation
Replace monolithic README with graduated structure: - Level 0: Ship in 5 minutes (3 files to edit, push) - Level 1: Understand the structure (skills, identity, CLAUDE.md) - Level 2: Advanced features (events, golden path, capability metadata) Reflects template overhaul: real example skills, graduated CI, dual identity layer, kishōtenketsu file structure. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c2717d658f
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| # 2. Edit three files | ||
| # construct.yaml → name, slug, description, author | ||
| # skills/example-simple/SKILL.md → your skill's instructions | ||
| # CLAUDE.md → your construct's identity |
There was a problem hiding this comment.
Restore command template customization in Level 0 quick start
The new Level 0 path says to edit only three files, but construct.yaml still maps the primary /quick-review command to commands/example-command.md, whose content is placeholder text that tells the agent to execute example-skill (a skill that is no longer declared). In practice, users who follow this onboarding flow can ship a construct whose main command runs stale instructions, so Level 0 should also require updating the command template (or point to a non-placeholder command file by default).
Useful? React with 👍 / 👎.
Summary
Context
Phase 2 of escape-velocity strategy. The README itself now embodies progressive disclosure — the principle the template teaches.
Test plan
🤖 Generated with Claude Code