Feature Development Workflow logo

Feature Development Workflow

Community
Firstp1ck
feature-development-workflow

Feature implementation workflow for authorized new capabilities. Use to classify a feature as lightweight or complex, resolve blocking decisions, and govern implementation, integration, review, reporting, and completion without applying complex gates to non-feature work.

Overview

PublisherFirstp1ck
Repositorypi-coding-agent-forge
Skill namefeature-development-workflow
Stars
93
Forks
22
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 Firstp1ck on GitHub. Read the source before you install it.

Installation

Install the Feature Development Workflow 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/Firstp1ck/pi-coding-agent-forge/tree/main/pi-skill-feature-development-workflow/skills/feature-development-workflow
  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/Firstp1ck/pi-coding-agent-forge.git /tmp/pi-coding-agent-forge
mkdir -p .claude/skills
cp -r /tmp/pi-coding-agent-forge/pi-skill-feature-development-workflow/skills/feature-development-workflow .claude/skills/feature-development-workflow
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Feature Development Workflow 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 Feature Development Workflow 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 Feature Development Workflow 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.

Feature Development Workflow

Use this skill for a user-authorized request to deliver a new capability. It preserves decision, implementation, review, and completion evidence while scaling the process to the feature's actual risk and scope. It is guidance, not a runtime guard: loading this skill does not install a package, enable an integration, change settings, or create hard enforcement.

When to Use

Use when the request adds a user-visible, product, system, interface, workflow, or integration capability and the agent is expected to implement it.

Do not use for:

  • Bug fixes that restore documented or intended existing behavior.
  • Refactors with no new capability.
  • Documentation-only, test-only, formatting-only, or dependency-maintenance work.
  • Planning, research, review, troubleshooting, incident response, operations, or a question without requested feature delivery.
  • Installation, enablement, publication, deployment, or configuration changes unless that action is separately authorized.

Should trigger

  • “Add a workspace export feature with authorization rules and a CSV download.”
  • “Implement a new notification preference channel and migrate existing defaults.”
  • “Create a small user-facing toggle that enables the requested display behavior.”
  • “Build this new integration endpoint and deliver it with tests.”

Should not trigger

  • “Fix the crash when an empty CSV is uploaded.”
  • “Refactor the cache module without changing behavior.”
  • “Add API documentation and unit tests for the existing endpoint.”
  • “Research options for notification providers before we choose one.”
  • “Review this pull request for security issues.”

Ambiguous requests

A request such as “add a tiny flag” may be a feature, a bug fix, or a configuration change. Inspect the affected behavior and user intent before routing. If it is feature delivery, classify it; otherwise use the workflow appropriate to the actual task.

Invocation Design

  • Invocation mode: model-invoked after explicit package enablement.
  • Leading concept: feature implementation workflow.
  • Route only when delivery of a new capability is requested; do not use this as a catch-all engineering process.
  • Creation of this package is inert. Review it before any later installation or enablement, which requires explicit authorization.

Inputs and Assumptions

Collect or establish:

  • User outcome, scope, constraints, and explicit authorization for side effects.
  • Relevant repository conventions, architecture, tests, deployment constraints, and existing interfaces.
  • Decisions that affect scope, architecture, interfaces, security, migration, compatibility, reliability, or rollout.
  • The available harness capabilities and any project planning/report conventions.

Do not ask again for a decision already answered by reliable project evidence. Do not begin implementation while a consequential blocking decision is unresolved.

Portable Workflow

  1. Preflight the required capability

    • Before entering a phase that mandates delegated implementation or independent review, verify that the selected harness can produce that outcome.
    • If the capability is unavailable, stop before that phase; report the unavailable capability and affected gate, then request an explicit user waiver or approved alternative. Do not silently lower a mandatory gate or claim completion.
    • Completion criterion: required capabilities are available, or a recorded explicit waiver/alternative exists.
  2. Inspect, decide, and classify

    • Inspect the codebase, conventions, constraints, and relevant documentation. Resolve only consequential decisions; group unresolved questions with a recommendation and trade-off.
    • Record the classification rationale. A feature is complex when it has at least two meaningful implementation slices, crosses components or contracts, requires migration or rollout work, has material security or reliability risk, or benefits from separate implementation and test/hardening ownership. Otherwise classify it as lightweight.
    • Validate any injected or inherited preliminary classification against repository evidence. Reclassify only when recorded material evidence contradicts it; if any complex criterion is met, keep or change the classification to complex.
    • Completion criterion: the feature class and rationale are recorded, and no blocking decision remains.
  3. Follow the lightweight path when classified lightweight

    • State success criteria, narrow scope, assumptions, and checks. Make the smallest safe implementation, run relevant checks, and report changed files, evidence, and residual risks.
    • A lightweight feature does not automatically require a canonical complex plan, two delegated implementation outcomes, two independent reviews, or an HTML report. Add only requirements that the user, repository, or actual risk independently requires.
    • Completion criterion: the requested capability is delivered within scope, relevant validation evidence is recorded, and remaining uncertainty is disclosed.
  4. Follow the complex path when classified complex

    • Read references/COMPLEX-FEATURE-CONTRACT.md before implementation. Create the canonical plan, preserve its workstream boundaries, and use its mandatory implementation, integration, review, report, waiver, and completion conditions.
    • Completion criterion: the reference's complex completion checklist is satisfied; otherwise report the feature as incomplete.
  5. Keep authorization and evidence current

    • Stop and resolve an unapproved product, scope, architecture, security, migration, compatibility, deployment, or interface decision before continuing.
    • Run affected and cross-workstream checks after integration. State checks not run, failures, waivers, and residual risks accurately.
    • Completion criterion: the final status distinguishes verified evidence from omissions, waivers, and risks.

Safety and Side Effects

  • Ask before external side effects, destructive actions, configuration changes, publication, installation, enablement, deployment, or overwriting user-authored work when authorization is unresolved.
  • Keep implementation ownership isolated; do not let concurrent writers modify the same shared files without a defined integration sequence.
  • Do not present model-invoked instructions as hard enforcement. Existing runtime controls, if any, remain separate.
  • Do not replace missing mandated evidence with a self-check, a planning pass, a repeated run, or a claim that the work is probably complete.

Scripts, References, and Dependencies

  • references/COMPLEX-FEATURE-CONTRACT.md contains the portable contract for complex features. Read it only after classification selects the complex path.
  • tests/test_skill_contract.py validates the routing, portability boundaries, and required policy terms using Python 3.10+ standard library only.
  • ../../tests/routing/feature-development-workflow.json contains repository routing examples relative to this skill directory.

No runtime package dependency is required. This package does not implement delegation, reviewer selection, runtime settings, or enforcement.

Verification

From the package root, run:

bash
npm test
npm pack --dry-run --json

For feature work, success means the classification, decisions, implementation, validation, and risks are evidence-backed. A complex feature is complete only under the contract reference; a lightweight feature is complete only under its scoped success criteria.

Pi Adapter

  • In Pi, read the active feature policy and the repository plan before beginning a governed feature workflow. The active policy owns feature outcomes and completion gates; delegation mechanics remain separate.
  • Before a Pi phase that requires delegated workers or reviewers, verify the delegation capability for the current harness.
  • Outside T3 Code, verify the subagent capability is active. Inside T3 Code, use its app-owned delegate_task capability, including harness-prefixed tool names, for implementation workers and independent reviewer runs. Discover current provider/model support with orchestrator_capabilities; do not infer usability merely from a catalog entry.
  • An available T3 delegate_task satisfies the subagent capability requirement; do not ask for a waiver or extra implementation approval solely because Pi's native subagent tool is absent. This changes the delegation mechanism, not required worker outcomes, review independence, authorization, or completion evidence.
  • If the selected harness cannot provide a mandatory delegation outcome, stop before that phase and ask the user for an explicit scoped waiver or approval of a named alternative.
  • Use Pi's repository-reading, file-editing, and validation tools to inspect evidence, make bounded changes, and verify commands. Use the Pi progress UI for multi-step work when available.
  • For a complex feature in Pi, place the canonical plan under the repository's planning convention (otherwise plans/planned/<feature-slug>.md), preserve one integration owner, and save the final report under the repository's report convention (otherwise reports/<feature-slug>.html). Use the html-report skill for that report when it is available.
  • Do not install this package or modify Pi settings without explicit user confirmation. A later approved local installation may use pi install <absolute-path-to-package>.

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 Feature Development Workflow AI skill do?

Feature implementation workflow for authorized new capabilities. Use to classify a feature as lightweight or complex, resolve blocking decisions, and govern implementation, integration, review, reporting, and completion without applying complex gates to non-feature work.

Why use Feature Development Workflow on TypingMind?

Because you install it once and use it with any model. Feature Development Workflow 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 Feature Development Workflow in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Firstp1ck/pi-coding-agent-forge/tree/main/pi-skill-feature-development-workflow/skills/feature-development-workflow. 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 Feature Development Workflow?

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 Feature Development Workflow?

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

Is the Feature Development Workflow AI skill free?

Yes. It is published on GitHub by Firstp1ck 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 👇