Analog Decompose logo

Analog Decompose

Community
Arcadia-1
analog-decompose

Architect Phase 1: decompose top-level analog spec into sub-block specs. Use when starting a new analog design to select architecture, allocate budgets, define sub-block specs, write testbenches, and create verification plans.

Overview

PublisherArcadia-1
Repositoryanalog-agents
Skill nameanalog-decompose
Stars
84
Forks
19
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by Arcadia-1 on GitHub. Read the source before you install it.

Installation

Install the Analog Decompose AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Arcadia-1/analog-agents/tree/main/skills/analog-decompose
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Arcadia-1/analog-agents.git /tmp/analog-agents
mkdir -p .claude/skills
cp -r /tmp/analog-agents/skills/analog-decompose .claude/skills/analog-decompose
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Analog Decompose in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Analog Decompose on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Analog Decompose is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

analog-decompose

Architect Phase 1: architecture decomposition. Takes a top-level analog spec and decomposes it into sub-block specs with budget allocation, testbenches, and verification plans. This is the first design activity in the pipeline.

Inputs

InputSourceRequired
Top-level spec.ymlWorking directoryyes
User constraints (architecture preference, area/power budget, etc.)Useryes (can be empty)
Wiki consult results/analog-wiki consultno — used if wiki/ has entries
Server name from servers.ymlConfigyes

Outputs

OutputLocationDescription
architecture.mdarchitect/architecture.mdArchitecture candidates, tradeoff analysis, selected architecture, block diagram
budget.mdarchitect/budget.mdPower/noise/timing budget allocation per sub-block
Sub-block spec.ymlblocks/<name>/spec.ymlDerived specs with derivation shown
Verification plansblocks/<name>/verification-plan.mdWhat to verify, how, and pass/fail criteria per level
Testbench netlistsblocks/<name>/testbench_<name>.scsStimulus, load, bias, analysis statements per sub-block

Effort

Reference shared-references/effort-contract.md for the full effort contract. Decomposition itself has no effort variation — the same thoroughness applies at all effort levels. Effort affects downstream skills (behavioral validation, iteration limits, review depth, etc.), not the architecture decomposition itself.

Wiki Interaction

At the start of decomposition, if the wiki/ directory contains entries, call:

/analog-wiki consult <block-type>

This returns relevant topologies, sizing strategies, anti-patterns, and historical project cases. Use these references to inform architecture selection and risk assessment. Wiki consult is advisory — the architect makes the final decision.

No wiki writes happen during decomposition.

Agent Template

Dispatch the architect agent using prompts/architect-prompt.md. Provide:

  • Path to top-level spec.yml
  • User constraints (architecture preference, area budget, power budget, etc.)
  • Server: value of role_mapping.architect in servers.yml (or default)
  • Wiki consult results (if available)

Checklist Responsibility

The architect is responsible for setting the checklists field in each sub-block's spec.yml. This field determines which pre-simulation checklists the verifier loads.

yaml
# In blocks/<name>/spec.yml (set by architect):
checklists: [common, amplifier, folded-cascode, differential]

If the checklists field is absent, the verifier falls back to keyword matching against the block field. Explicit assignment is preferred — see shared-references/checklist-schema.md for the keyword-to-checklist mapping.

Architecture Decomposition Rules

  • Every sub-block must have a clear, testable interface — defined input/output signals with impedance/swing/timing specs
  • No circular dependencies — block A's spec cannot depend on block B's implementation if B depends on A
  • Budget must close — if individual block budgets don't sum to meet the top-level spec, the architecture is invalid. Iterate before proceeding.
  • Flag risks at block boundaries — e.g., "comparator kickback will disturb DAC top plate — need isolation switch or timing guard"

Spec Derivation Example (SAR ADC)

Top-level: 8-bit, 100MS/s, ENOB >= 7.5

Noise budget (thermal):
  Total kT/C noise < LSB/2 / 6.6sigma -> C_sample > 4kT / (V_LSB/6.6)^2
  Allocate: 70% to sampling cap, 20% to comparator, 10% to reference

Timing budget (100MS/s -> 10ns per conversion):
  Sampling phase: 3ns (30%)
  8 SAR cycles: 7ns -> 875ps per bit
    - DAC settling: 400ps
    - Comparator decision: 400ps
    - Logic delay: 75ps

Power budget (e.g., 2mW total):
  Comparator: 0.8mW (40%)
  DAC switching: 0.4mW (20%)
  SAR logic: 0.2mW (10%)
  Reference/bias: 0.6mW (30%)

Each sub-block spec.yml must show the derivation from top-level requirements, not just state the number. The derivation is what makes the spec auditable.

Testbench Writing Guidelines

The architect owns all testbenches. When writing testbenches, keep these common pitfalls in mind — the verifier will check for them before simulating:

  • PSRR: supply ripple must be applied with the input properly biased and output loaded; measure at the output, not an internal node
  • CMRR: common-mode input must sweep while differential input is zero; watch for CMFB loop interaction
  • Phase margin: loop must be broken at the correct point with correct replica loading; Middlebrook or STB analysis preferred over open-loop if feedback is complex
  • Noise: check that the noise bandwidth and integration limits match the spec definition; spot noise vs integrated noise are different specs
  • Settling time: input step must be realistic (not larger than linear range); measure to the correct accuracy band (e.g., 0.1% vs 1 LSB)
  • Power: measure at steady state, not during startup transient

If the verifier rejects a testbench, fix the issue and resubmit. This is not a design iteration — it is quality control on the architect's own work.

Handoff Contract

Reference shared-references/handoff-contracts.md for the full handoff payload requirements.

Phase 1 Inputs (Top-level spec -> Architect)

FieldRequired
Top-level spec.yml pathyes
User constraintsyes (can be empty)
Server name from servers.ymlyes

Phase 1 Outputs (Architect returns)

ArtifactRequired
architecture.md with candidates, tradeoffs, selected architectureyes
budget.md with power/noise/timing budgets that closeyes
blocks/<name>/spec.yml per sub-block with derived specsyes
blocks/<name>/verification-plan.md per sub-blockyes
blocks/<name>/testbench_<name>.scs per sub-blockyes

Handoff Acceptance Criteria

Phase 1 is complete when:

  • architecture.md documents candidates, tradeoffs, and selected architecture
  • Every sub-block has blocks/<name>/spec.yml with derived specs and derivation shown
  • Every sub-block has blocks/<name>/verification-plan.md defining what to verify and how
  • budget.md shows power/noise/timing budgets close (sum meets top-level)
  • Every sub-block spec.yml has the checklists field set
  • No transistor-level design has been done (that is the designer's job)

Scope

This skill covers Phase 1 only. Phase 2 (behavioral validation) is handled by /analog-behavioral. Phase 3 (integration verification) is handled by /analog-integrate.

Frequently asked questions

What does the Analog Decompose AI skill do?

Architect Phase 1: decompose top-level analog spec into sub-block specs. Use when starting a new analog design to select architecture, allocate budgets, define sub-block specs, write testbenches, and create verification plans.

Why use Analog Decompose on TypingMind?

Because you install it once and use it with any model. Analog Decompose is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Analog Decompose in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Arcadia-1/analog-agents/tree/main/skills/analog-decompose. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Analog Decompose?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Analog Decompose?

As many as you like. As long as a model supports skills, you can use Analog Decompose with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Analog Decompose AI skill free?

It is published on GitHub by Arcadia-1. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇