Skip to content
Adaptive Skills Docs
Esc
↑↓navigate↵open⌘Jpreview
On this page

Claude Consumer Setup

Reference documentation for Claude Consumer Setup in Adaptive Skills.

Goal

Use Adaptive Skills in a Claude-oriented workflow without pretending that v0 supports symmetric auto-installation.

In v0, Claude projection is intentionally selective. That means the repository helps you decide what to reuse, but it does not try to force every skill into a Claude-native installation shape.

Start with skill selection, not projection symmetry. A skill can be useful in Claude by reference before it needs a Claude-specific wrapper.

Current Claude modes

The projection registry uses three modes:

  • link-only — keep the canon in this repository and point local docs or workflows to it
  • manual — adapt the skill deliberately for the Claude workflow in your consumer project
  • available-not-default — the skill is still available for Claude use, but it is not part of the current default operating path

You can inspect the current state with:

python3 scripts/report_projection_status.py
python3 scripts/project_to_claude.py

1. Choose the skill from the task

Read:

  • docs/how-to-use-a-skill.md
  • docs/consumer-adoption.md

Then decide:

  • what task or lane Claude is supporting
  • which skill fits the dominant need
  • whether the canonical skill is enough by reference
  • whether the consumer project needs a local wrapper

2. Start from the canon, not from duplication

Read the selected skill in this repository first. If the skill is enough as-is, use link-only behavior in your local project notes or workflow docs.

For example, premortem can often be used by linking to the canonical skill and asking Claude to follow its core moves. A wrapper is only needed if the team has a repeated Claude-specific planning ritual.

3. Adapt only the skills that really need Claude-specific handling

Good manual candidates are skills where:

  • collaboration style matters
  • the work is heavily conversational
  • the consumer team already has a Claude-specific ritual or prompt style

Examples in v0:

  • ux-strategy
  • ux-provocation
  • heuristic-audit
  • ux-writing
  • business-design
  • triad-check

4. Treat non-default skills as available, not blocked

If a skill is marked available-not-default, it is still usable with Claude. It simply means the repository is not treating it as part of the current default Claude path.

Practical setup model

A good Claude-oriented consumer setup usually looks like this:

  • Adaptive Skills repo -> canonical skill definition
  • consumer project docs -> local notes about when Claude should use or adapt those skills
  • consumer project Claude setup -> only the small subset that truly benefits from a Claude-specific wrapper

Recommended local note shape:

## Adaptive Skills used in this project

- `workflow` -> use by reference for bounded work slices.
- `feature-planning` -> use by reference before implementation.
- `premortem` -> use by reference for consequential plans; do not run by default.
- `ux-writing` -> local Claude wrapper allowed for recurring conversational copy reviews.

Suggested local pattern

A consumer project can keep a short local note such as:

  • use workflow by link/reference only
  • adapt ux-writing manually for Claude if the team already reviews copy conversationally
  • keep testing as available-not-default until there is a real Claude-native reason to wrap it
  • keep premortem as reference-first unless consequential planning becomes a recurring Claude workflow

When to choose manual adaptation

Use manual adaptation when:

  • the team needs a Claude-specific prompt style
  • the local workflow changes the reading order or output ritual
  • the skill needs a smaller wrapper for a recurring conversational loop

Do not choose manual adaptation when:

  • you are only chasing symmetry with Codex
  • the canonical skill already works by reference
  • the local project has not actually used the skill yet
  • the skill is only needed occasionally and can be linked from the canon

What not to do

  • do not mirror the entire library into Claude on day one
  • do not create Claude wrappers before selecting the skills by task need
  • do not rewrite the canon just to fit a local Claude wrapper
  • do not treat available-not-default as missing support
  • do not collapse local Claude rituals into the portable library

Relationship to AletheIA

If you use AletheIA, keep the same split:

  • Adaptive Skills -> micro execution guidance
  • AletheIA -> macro framing, gates, continuity, handoff

Claude-specific wrapping stays local to the consumer project unless it becomes broadly reusable.

Was this page helpful?