Radioactive logo

Radioactive

Community
devton
radioactive

Full-cycle 11-phase development workflow blueprint. Chains request classification, discovery, UX design, planning, blueprint creation, execution, testing, quality review, iterative fix loop, blueprint updates, and changelog generation.

Overview

Publisherdevton
Repositoryagentic-workflow-blueprint
Skill nameradioactive
Stars
73
Forks
4
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 devton on GitHub. Read the source before you install it.

Installation

Install the Radioactive 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/devton/agentic-workflow-blueprint/tree/main/workflows/radioactive
  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/devton/agentic-workflow-blueprint.git /tmp/agentic-workflow-blueprint
mkdir -p .claude/skills
cp -r /tmp/agentic-workflow-blueprint/workflows/radioactive .claude/skills/radioactive
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Radioactive 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 Radioactive 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 Radioactive 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.

agentic-workflows-blueprint.workflow.radioactive

Goal

Provide a portable, end-to-end 11-phase development lifecycle workflow that enforces strict quality gates from initial problem discovery to final review, fix iteration, institutional memory update, and changelog generation.


Scope

  • Applies to: Non-trivial features, architectural refactors, or cross-layer implementations (e.g. backend + API + frontend) across any project or tech stack.
  • Does not cover: Single-file trivial bug fixes, pure documentation edits (use document), or isolated test additions.

Triggers

  • "Run radioactive lifecycle"
  • "/radioactive [feature description]"
  • "Full-cycle feature development"
  • "Implement feature with end-to-end quality gate"

Inputs

  • featureDescription: Short text describing the feature or task (optional; if missing, triggers Phase 2 discovery).
  • projectSlug: Short identifier for the repo (e.g., apix, my-backend).
  • baseBranch: Target integration branch (default: main or develop).
  • techStack: Stack details (e.g., Ruby on Rails + PostgreSQL, Next.js + TypeScript).
  • existingRootDoc: Main instruction file (AGENTS.md, CLAUDE.md, etc.).
  • uiInvolved: Boolean flag indicating if frontend/UI changes are required.
  • maxFixAttempts: Maximum iterations for the Phase 9 fix loop (default: 3).

Invariants (Guardrails)

  1. Sequential Phase Pipeline: All 11 phases must execute in exact numerical order. No phase skipping.
  2. Abstract Entrypoint Fallback:
    • If a primary project skill exists (e.g. skills/<projectSlug>/SKILL.md or .agents/skills/<projectSlug>/SKILL.md or AGENTS.md): Read it, classify request, and enforce project-specific hard rules and routing.
    • If NO primary project skill exists: Trigger /workflow-blueprint init (or plan-to-blueprint) to scaffold baseline project context and routing before proceeding.
  3. Socratic Discovery Gate: Phase 2 must present discovery questions and wait for explicit user response/confirmation before Phase 3/4.
  4. Plan Approval Gate: Phase 4 must generate a structured task breakdown at docs/tasks/task.<reference>.md and obtain explicit user approval before Phase 5.
  5. Executable Blueprint Contract: Phase 5 must scaffold a reusable workflow contract at skills/<projectSlug>/workflows/<feature-slug>/SKILL.md.
  6. Pass/Fail Quality Gate: Phase 7 specs must pass 100%. Phase 9 fix loop must resolve all 🔴 (Blockers) and 🟠 (High) findings before declaration of completion.
  7. Institutional Memory: Phase 10 must capture newly discovered patterns into workflow contracts and update project documentation.
  8. Operational Safety for Infra Tasks: When the feature affects network, infrastructure, IaC, or OS baselines, execution must include explicit rollback, change window alignment, and post-change health validation.

Procedure

Phase 1  — CLASSIFY & INIT  → Project skill or /workflow-blueprint init
Phase 2  — DISCOVER         → /brainstorming (Socratic questions & decision log)
Phase 3  — UX DESIGN        → /ui-ux-pro-max or frontend-design (if UI involved)
Phase 4  — PLAN             → /plan-writing (structured task breakdown in docs/tasks/task.<reference>.md)
Phase 5  — BLUEPRINT        → /workflow-blueprint (scaffold feature contract)
Phase 6  — EXECUTE          → Implement tasks sequentially with evidence-based edits
Phase 7  — TESTS            → Project test runner (specs/tests for all changed code)
Phase 8  — REVIEW           → /thermo-nuclear-code-quality-review or /review
Phase 9  — FIX LOOP         → /thermo-fix × up to 3 (resolve 🔴→🟠→🟡, verify build, re-review)
Phase 10 — BLUEPRINTS       → /workflow-blueprint (update skills & internal contracts)
Phase 11 — CHANGELOG        → /changelog-generator or /changelog (user-facing release note)
Phase 1 — CLASSIFY & INIT
  1. Check for the existence of the primary project entry skill (skills/<projectSlug>/SKILL.md, .agents/skills/<projectSlug>/SKILL.md, or AGENTS.md).
  2. Branch logic:
    • Option A (Project skill exists): Load the skill, classify the task using the project's routing matrix, and extract hard constraints and domain boundaries.
    • Option B (Project skill does NOT exist): Initialize /workflow-blueprint (run scaffolding / plan-to-blueprint flow) to analyze the project structure, detect projectSlug and techStack, and establish a baseline entry skill.
  3. Output a 1-paragraph context summary and the governing workflow contract.
Phase 2 — DISCOVER with /brainstorming
  1. Load brainstorming skill (or run Socratic discovery protocol).
  2. Ask 3–5 strategic questions covering scope boundaries, data models, security surface, integration points, and edge cases.
  3. Wait for user response. Do NOT proceed to Phase 3/4 until questions are answered or user explicitly approves.
  4. Summarize into a decision log appended to conversation context.
Phase 3 — UX DESIGN (Conditional)
  1. Check condition: Run only if uiInvolved is true (or feature touches views, templates, or visual components).
  2. Load UI design skill (ui-ux-pro-max or frontend-design).
  3. Generate design tokens (color palette, CSS variables, typography) and responsive layout rules (using mobile-first min-* breakpoints).
  4. Append UX design rules to the decision log.
Phase 4 — PLAN with /plan-writing
  1. Load plan-writing skill (or task planning methodology).
  2. Produce a structured task specification at docs/tasks/task.<reference>.md containing context, tasks with file targets, explicit dependencies, and verification criteria.
  3. Present plan to user and gate on explicit approval.
Phase 5 — BLUEPRINT with /workflow-blueprint
  1. Load workflow-blueprint (or plan-to-blueprint).
  2. Transform the approved plan into an executable workflow contract saved at skills/<projectSlug>/workflows/<feature-slug>/SKILL.md.
  3. Wire the new feature workflow into the project's routing matrix.
Phase 6 — EXECUTE
  1. Execute tasks in docs/tasks/task.<reference>.md sequentially.
  2. Read files before editing; apply minimal, targeted edits.
  3. Track progress by marking [/] (in progress) and [x] (completed) in docs/tasks/task.<reference>.md.
  4. For infra-impacting tasks, execute pre-change checks and capture rollback checkpoints before mutating state.
Phase 7 — TESTS
  1. Run the project's test runner (e.g. rspec-rails, vitest, pytest, or testing-patterns).
  2. Write unit/integration/system tests for all newly created or modified logic.
  3. Iterate until 100% of specs pass.
  4. For operational tasks, run infrastructure checks (connectivity, policy checks, service health, drift validation) and archive outputs.
Phase 8 — REVIEW with Code Quality Engine
  1. Run code review (thermo-nuclear-code-quality-review or review skill) against all files modified in the session.
  2. Output a severity-ranked findings table:
    • 🔴 Blocker: Must fix before release.
    • 🟠 High: Fix in current cycle.
    • 🟡 Medium: Fix if trivial or record.
    • 🟢 Low: Informational.
  3. Append findings table to walkthrough.md artifact.
Phase 9 — FIX LOOP (up to maxFixAttempts ×)
  1. Iterate up to maxFixAttempts times:
    • Apply fixes for 🔴 Blocker and 🟠 High findings.
    • Run build verification command (bun run build, rails runner, npm run test, etc.).
    • Re-review modified files.
    • Stop loop early if zero 🔴 / 🟠 findings remain (verdict: APPROVED).
  2. Update docs/tasks/task.<reference>.md and walkthrough.md with final verdict.
  3. Create a clean git commit on the local working branch.
Phase 10 — BLUEPRINTS Update
  1. Review discoveries made during execution.
  2. Update existing project skills with new patterns, constraints, or schema details.
  3. Scaffold new internal workflow blueprints if repeatable operations were introduced.
  4. Update root doc (AGENTS.md) if core system rules changed.
Phase 11 — CHANGELOG Generation
  1. Load changelog-generator (or changelog skill).
  2. Analyze session git commits against base branch.
  3. Generate customer-facing release notes categorized by Features, Fixes, and Improvements.
  4. Append changelog entry to walkthrough.md.

Outputs

  • docs/tasks/task.<reference>.md: Executed task breakdown with completion status.
  • walkthrough.md: Decision log, UX specs, review tables, fix history, and user changelog.
  • skills/<projectSlug>/workflows/<feature-slug>/SKILL.md: Executable contract for the feature.
  • Updated project skills and routing matrix.
  • Clean git commit on local branch.

Review gate

  • Phase 1 project skill loaded or initialized via /workflow-blueprint.
  • Phase 2 discovery questions answered by user.
  • Phase 3 UX design system generated (if UI feature).
  • Phase 4 plan approved by user.
  • Phase 5 workflow contract created & registered in project routing matrix.
  • Phase 6 execution complete.
  • Phase 7 all test suites passing.
  • Phase 9 zero Blocker or High findings remaining.
  • Phase 10 skills/blueprints updated with new knowledge.
  • Phase 11 user-facing changelog generated.

References

  • ../../SKILL.md
  • ../plan-to-blueprint/SKILL.md
  • ../review/SKILL.md
  • ../changelog/SKILL.md

Frequently asked questions

What does the Radioactive AI skill do?

Full-cycle 11-phase development workflow blueprint. Chains request classification, discovery, UX design, planning, blueprint creation, execution, testing, quality review, iterative fix loop, blueprint updates, and changelog generation.

Why use Radioactive on TypingMind?

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

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/devton/agentic-workflow-blueprint/tree/main/workflows/radioactive. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Radioactive?

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 Radioactive?

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

Is the Radioactive AI skill free?

It is published on GitHub by devton. 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 👇