ship-adapter — Claude Code skill: adapters and installation
The ship-adapter Claude Code skill turns schemas and API contracts into adapters. It runs for database adapters, API client wrappers, interface adapters, and data transformers.
MIT ★ 26 updated 2026-11-16
npx skills add wrg32786/aigent-os --skill ship-adapter -a claude-codeWhat ship-adapter does
The skill writes code that translates between two known interfaces, such as a database schema and an API response type. It reads both definitions and searches the project for adapter patterns before matching the existing error-handling style. Both sides of the contract are read before any code is written.
Null and optional fields get explicit handling, and unusual field mappings receive comments. The result includes a record of changed and untouched files, unresolved mapping issues, assumptions, and tradeoffs. Its scope ends at the adapter; infrastructure work goes back to the composing agent.
What it helps with
- translate data between two schemas
- wrap an API with a client interface
- map database fields to a response type
- handle null and optional fields in a transformation
Example prompts for the agent
- Write an adapter from the
UserRecordschema to theUserProfileresponse type. - Wrap this API in a client adapter that follows the project’s existing patterns.
- Transform this schema to the target interface without dropping fields.
How to install ship-adapter
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill ship-adapter -a claude-code2Manually: put the files in place
~/.claude/skills/ship-adapter/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/ship-adapter/
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill ship-adapter -a codex2Manually: put the files in place
.agents/skills/ship-adapter/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill ship-adapter -a cursor2Manually: put the files in place
.cursor/skills/ship-adapter/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install wrg32786/aigent-os skills/ship-adapter2Manually: put the files in place
.github/skills/ship-adapter/SKILL.md1Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill ship-adapter -a gemini-cli2Platform CLI
gemini skills install https://github.com/wrg32786/aigent-os --path skills/ship-adapterManually: put the files in place
.devin/skills/ship-adapter/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add wrg32786/aigent-os --skill ship-adapter -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about ship-adapter
How do I install ship-adapter in Claude Code?
Run npx skills add wrg32786/aigent-os --skill ship-adapter -a claude-code. The skill file is at ~/.claude/skills/ship-adapter/SKILL.md.
What does ship-adapter do?
It writes an adapter between two interfaces or schemas and follows patterns found in the project. It returns a record of the files changed and decisions made.
What happens to fields that have no mapping?
The skill does not silently discard them; it adds a comment for each unmapped field. It also comments on mappings that need explanation.
Source text
SKILL.md · MIT
--- name: ship-adapter agent: lyra description: "Port schemas, interfaces, and API contracts into adapters. Use when writing a DB adapter, API client wrapper, interface adapter, or data transformer." allowed-tools: Read, Write, Edit, Bash, Grep, Glob, Agent user-invocable: true triggers: - ship adapter - port adapter - write adapter - DB adapter - API adapter - data transformer - schema migration - interface adapter - wrap this API - transform this schema --- # Ship Adapter You are Lyra shipping a bounded adapter — a translation layer between two data shapes, APIs, or storage systems. ## What this skill does Write an adapter that translates between two known interfaces. Inputs: source shape + target shape + any behavioral constraints. Output: working adapter code + honesty ledger. ## Protocol ### Step 1: Confirm the contract State in one sentence what is being adapted. Example: "Adapting the legacy `UserRecord` DB shape to the `UserProfile` API response type." Identify: - Source shape (where data comes from) - Target shape (what consumers expect) - Edge cases or nullability constraints called out in the spec If ambiguous, ask ONE question. Then build. ### Step 2: Read both sides - Read the source type/schema definition - Read the target type/schema definition - Grep for existing adapter patterns in the codebase to match style ### Step 3: Write the adapter - Handle nulls and optional fields explicitly — no silent drops - Add a comment for any non-obvious field mapping - Match the codebase's existing error handling pattern ### Step 4: Return honesty ledger ``` ## Honesty Ledger **Changed:** <files created or modified> **Untouched:** <files read but not changed> **Noticed-not-fixed:** <mapping issues seen but out of scope> **Residual uncertainty:** <fields that required a guess> **Tradeoffs:** <decisions made — e.g., lossy vs lossless, null coalescing strategy> **Stopped-short:** <anything the spec implied but wasn't confirmed> ``` ## Constraints - Read source and target types before writing a single line. - No silent field drops — every unmapped field gets a comment. - Scope boundary is the adapter itself. Infrastructure changes go back to the composer.
© wrg32786 · MIT · wrg32786/aigent-os