skills.rest
biocypher
biocypherOFFICIAL
Explore all skills by this author →
WebsiteGitHub

karenina-adapter-design

Generates a design spec for a new karenina adapter from a gathered context document.

13|3|Updated Jun 27, 2025
Search Tags:#agent sdk#spec generation#concept mapping#llm integration#evaluation framework#adapter design#karenina
One-click install
npx skills add https://github.com/biocypher/karenina --skill karenina-adapter-design-biocypher
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: karenina-adapter-design
Source: https://github.com/biocypher/karenina/tree/main/skills/karenina-adapter-design
Command: npx skills add https://github.com/biocypher/karenina --skill karenina-adapter-design-biocypher

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Designing a new adapter for the karenina evaluation framework requires mapping many port concepts (AgentPort, LLMPort, ParserPort) to an unfamiliar SDK, and missing details like usage aggregation or timeout handling leads to silent bugs. This Skill turns a gathered context document into a complete, structured adapter design spec. ## Core Features & Use Cases - Concept Mapping Tables: Produces full mapping tables covering AgentPort, LLMPort, and ParserPort concepts against the target SDK equivalents. - Pitfall Coverage: Explicitly addresses known failure areas such as filesystem backends, recursion limits, multi-turn usage aggregation, streaming timeouts, and structured output return types. - Spec Document Generation: Writes an 18-section design spec to docs/adapters/<name>-adapter-design.md, including registration code, core flow, error mapping, and success criteria. - Use Case: After running the context-gathering phase for a new agent SDK, use this Skill to produce the design spec that the implementation phase will follow. ## Quick Start Design a karenina adapter for the SDK described in docs/adapters/my-sdk-context.md and save the spec.

Frequently Asked Questions about karenina-adapter-design

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design a new karenina adapter for an agent SDK?▼

Run this Skill after creating a context document in Phase 1. It reads docs/adapters/<name>-context.md, maps all karenina port concepts to SDK equivalents, and writes a full design spec to docs/adapters/<name>-adapter-design.md.

What ports must a karenina adapter design cover?▼

The design must cover all three ports: AgentPort for multi-turn agent loops, LLMPort for single-turn calls and streaming, and ParserPort for structured Pydantic extraction. Each port requires a complete concept mapping table.

What is the prerequisite for running the adapter design phase?▼

A context document must exist at docs/adapters/<name>-context.md, produced by the karenina-adapter-gather-context Skill in Phase 1. It contains SDK capabilities, preliminary concept mappings, and resolved design questions.

Which common adapter bugs does the design spec address?▼

The spec explicitly addresses virtual filesystem backends, lost usage data in structured output, recursion limit handling, multi-turn token aggregation, turn limit mapping, no-tools fallbacks, timeout coverage on all paths, and partial trace recovery on agent timeouts.

How are agent_tier and adapter characteristics decided?▼

The Skill sets agent_tier to deep_agent if the SDK manages its own tool loop, or tool_loop if karenina orchestrates turns. It also determines supports_mcp, supports_tools, and fallback_interface based on the gathered SDK capabilities.

Related Skills

Explore More Software Engineering Skills→
KnoWhizKnoWhiz
6

mcp-builder

Guide developers in creating MCP servers with Python FastMCP and Node TypeScript SDK.

Official
Advanced
windingriverholdingswindingriverholdings

create-agent-adapter

Define Paperclip agent adapter packages with required exports and registration conventions.

Official
Advanced
mattjeffersonmattjefferson

brainstorming

Resolve unclear requirements through a structured phase-based brainstorming process.

Community
Intermediate
real-simple-labsreal-simple-labs
44

scriptwriting

Write ad scripts matched to brand, audience, awareness stage, and format.

Official
Advanced