Thanks for your interest in contributing. Here's how to get involved.
- Fork and clone the repo
- Install Claude Code
- Install the plugin locally:
claude plugin add /path/to/your/clone - Run
/climbto test the onboarding flow
bots/
.claude-plugin/ # Repo-level plugin discovery
marketplace.json # Registry of available plugins
plugin.json # Repo metadata
plugins/
ladder/ # The Ladder personal growth plugin
.claude-plugin/
plugin.json # Plugin metadata
commands/
climb.md # /climb entry point (onboarding + daily rhythm)
references/
enneagram-seed-patterns.md # Seeded patterns for all 9 types
skills/
goals/ # Goal creation and revision
references/ # Goal file format template, research foundation
reflections/ # Weekly structured reviews
references/ # Reflection file format template
accountability/ # Commitment accountability
coach/ # Framework-based coaching (9 topic areas)
references/ # 9 curated reference files
Each skill is a markdown file with YAML frontmatter (name + description) and structured instructions. The description field controls when the skill triggers. The body contains the steps the model follows.
Reference files under references/ are loaded by explicit Read instructions in the skill. If you add a reference file, you must add a corresponding Read step — a file that nothing reads is decoration.
The coach skill has 9 topic areas. To add a new one:
- Create
skills/coach/references/your-topic.md - Follow the existing format: topic header, intro line, author sections with "Apply when:" guidance
- Add the topic to the mapping table in
skills/coach/SKILL.md - Use published, attributed frameworks only — name the author and source
The seed patterns at commands/references/enneagram-seed-patterns.md are hypotheses based on type psychology. If you have Enneagram expertise and see:
- A pattern assigned to the wrong type
- A missing pattern that's core to a type's psychology
- A response approach that would backfire for the stated type
Open an issue or PR with your reasoning.
The Enneagram-to-calibration-defaults table in commands/climb.md maps each type to 7 settings. If a default doesn't fit (e.g., a type's accountability intensity is too high or too low), explain why in your PR.
If you find:
- A reference file that nothing reads
- A file path that doesn't resolve
- Inconsistent terminology between skills
- A quality gate that doesn't actually check what it claims
These are high-priority fixes. Please include evidence (grep output, file paths, line numbers).
- No proprietary frameworks. All coach reference content must be from published, publicly available sources with author attribution.
- No clinical language. This is a personal growth tool, not therapy. Avoid clinical diagnoses, pathological framing, or language that implies the user has a disorder.
- Test the full flow. If you change onboarding, run through it as a new user. If you change a skill, invoke it and verify it reads the files it claims to read.
- Keep skills self-contained. Every skill must resolve
kb_pathindependently (Step 0: Load Config). Don't assume another skill already loaded the data.
Open a GitHub issue with:
- What you expected to happen
- What actually happened
- Which skill/command was involved
- Your Enneagram type and calibration settings (if relevant to the issue)
Be kind. This tool deals with personal growth — contributions should reflect that spirit.