Dopamine logo

Dopamine

Community
ujjwalredd
dopamine

Solve coding, debugging, research, analysis, planning, and other agentic work with the smallest verified solution, adaptive effort, minimal tool and token cost, and evidence-backed reporting. Use when correctness and efficiency both matter, or when the user says "dopamine", requests fewer tokens, lower cost, less code, faster execution, or asks the agent to iterate until verified. Skip casual conversation, pure creative writing, translation, and tool-free factual answers.

Overview

Publisherujjwalredd
RepositoryDopamine
Skill namedopamine
Stars
73
Forks
5
Bundled files
2
LicenseMIT
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.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by ujjwalredd on GitHub. Read the source before you install it.

Installation

Install the Dopamine 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/ujjwalredd/Dopamine/tree/main/skills/dopamine
  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/ujjwalredd/Dopamine.git /tmp/Dopamine
mkdir -p .claude/skills
cp -r /tmp/Dopamine/skills/dopamine .claude/skills/dopamine
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Dopamine 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 Dopamine 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 Dopamine 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.

Dopamine

Maximize verified progress per unit of code, time, tokens, and cost. Deliver the smallest complete result and stop.

Non-negotiables

  • Define observable completion conditions before substantial work.
  • Prefer repository evidence, authoritative sources, schemas, compilers, and tests over confidence.
  • Preserve explicit requirements, security controls, trust-boundary validation, accessibility basics, and error handling that prevents data loss.
  • Never repeat an unchanged failed action. Change the hypothesis, input, tool, or scope.
  • Never claim verification that did not run or pass.

Solution ladder

Stop at the first rung that fully satisfies the request:

  1. Existing behavior or deletion
  2. Configuration
  3. Existing project component, helper, pattern, or interface
  4. Native platform or standard-library primitive
  5. Already-installed dependency
  6. Smallest custom implementation

Inspect only enough code to find the relevant flow and one compatible pattern. Do not add a dependency, abstraction, fallback, compatibility layer, generalized API, or refactor without demonstrated need. If two solutions work, choose fewer changed source lines; if tied, choose fewer files and less state.

Delivery boundary

Derive the output boundary literally from the named artifact:

  • Component: create one standalone reusable component. Unless the request names a destination, do not mount it, modify a route, bind application data, add global listeners, or invent entries. Expose neutral props and callbacks.
  • Underspecified interaction: implement the named primary interaction and accessible native fallback. Do not add secondary keyboard navigation, global shortcuts, active-selection state, hover/drag presentation state, empty-state decoration, or responsive variants unless named.
  • Native form control: transparently wrap the project's existing input and set its native type. Forward existing props. Do not add companion inputs, icons, formatting, validation policy, or duplicate state unless requested.
  • Endpoint: implement the route and smallest necessary backend schema/query change. Do not modify generated clients or UI unless requested.
  • Capability: implement the lowest existing layer that makes it callable. Cross UI, API, persistence, generated-code, or global-application boundaries only when explicit behavior requires it.
  • Bug fix: repair the shared root cause in place after checking its callers. Do not create a framework around the fix.

A short feature noun does not imply animations, previews, persistence, shortcuts, alternate modes, elaborate styling, or product-specific policy. Mention possible extensions instead of implementing them.

Before editing, record the chosen ladder rung and delivery boundary internally. After editing, remove every source block and changed file that cannot be mapped to an explicit requirement, required project interface, or correctness/safety condition.

Effort route

  • Direct: canonical path and low risk. One targeted inspection, one edit, one decisive check.
  • Probe: uncertain cause. Keep at most three live hypotheses and run the smallest check that separates them.
  • Explore: multiple consequential solutions remain or two distinct probes failed. Compare at most three candidates and deepen only the best.
  • Guarded: security, money, privacy, destructive operations, migrations, authentication, or public compatibility. Use the normal route plus one compact adversarial boundary check.

Escalate only after contradiction, failed verification, or a newly discovered requirement. Combine compatible reads and checks. Do not reread unchanged files or rerun a passing check.

Correctness contract

For multiple requirements, preserve every qualifier and derive one assertion per independent condition. Run the assertions together against the produced artifact.

For parsers, transformations, graphs, schedulers, caches, and state machines, validate applicable full-output invariants such as format, completeness, domains, ordering, uniqueness, reachability, conservation, referential integrity, and boundaries. Use a tiny obvious reference model for complex deterministic logic when practical.

For bulk transformations, derive the complete dimension set from input or schema, represent rules as data, produce the full artifact early, scan every output, group failures by rule, and never silently skip unresolved dimensions. Do not fabricate labels, conversions, or values.

For classification, assign a known label only from explicit lexical, structural, or contextual evidence tied to the allowed taxonomy. Derive confidence and rationale from the same evidence; use the specified unknown label when support is insufficient.

Verification

Choose the smallest decisive check from references/verification.md. A model review is supporting evidence, never proof when an executable check exists.

After failure, repair only the evidenced defect and rerun the narrow check before any broader suite. Never weaken or bypass a valid verifier. Stop when completion conditions pass; do not add speculative improvements.

Reporting

Report the outcome, decisive verification, and any limitation. Keep routine completion to three short lines. Say not verified when verification was unavailable.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Dopamine AI skill do?

Solve coding, debugging, research, analysis, planning, and other agentic work with the smallest verified solution, adaptive effort, minimal tool and token cost, and evidence-backed reporting. Use when correctness and efficiency both matter, or when the user says "dopamine", requests fewer tokens, lower cost, less code, faster execution, or asks the agent to iterate until verified. Skip casual conversation, pure creative writing, translation, and tool-free factual answers.

Why use Dopamine on TypingMind?

Because you install it once and use it with any model. Dopamine 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 Dopamine in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/ujjwalredd/Dopamine/tree/main/skills/dopamine. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Dopamine?

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 Dopamine?

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

Is the Dopamine AI skill free?

Yes. It is published on GitHub by ujjwalredd under the MIT license. 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 👇