AGENT SKILL

smart-init

Conversational onboarding for A Team. Detects ROADMAP.md, extracts context, generates INIT.md without requiring technical knowledge. Invoked automatically by the orchestrator when INIT.md is missing.

License 许可
Apache-2.0
最近更新
2026-11-13
Stars
★ 1.6k
Forks
483

01 前言 Front Matter

name
smart-init
description
Conversational onboarding for A Team. Detects ROADMAP.md, extracts context, generates INIT.md without requiring technical knowledge. Invoked automatically by the orchestrator when INIT.md is missing.

03 SKILL.md 正文

Smart Init — Conversational Onboarding

Trigger

This skill is invoked by the orchestrator when /orchestrate init is called and no INIT.md exists.

Detection Sequence

Check in this order:

  1. INIT.md exists → stop, use current init flow unchanged
  2. ROADMAP.md exists → Path A (extract from ROADMAP)
  3. ROADMAP_*.md exists (e.g. ROADMAP_icd10.md) → Path A (extract from that file)
  4. Neither exists → Path B (full 5-question interview)

Path A — ROADMAP Exists

Read the ROADMAP file and extract:

Look for in ROADMAP Maps to INIT.md field
Project name in H1 or title Project name
Description paragraph Project overview
Stack table or technology mentions Languages & stack
"Princípios não-negociáveis" or "Non-negotiable" section Immutable rules
"Próximo" / "Next" / "Roadmap" items Active work context
Compliance mentions (GDPR, local-first, privacy, HIPAA) Compliance scope
"Feito" / "Done" section Existing coverage

After extraction, ask only one question (the only thing a ROADMAP cannot tell you):

"Que ferramentas de IA estás a usar para escrever código neste projecto?" → A) Claude Code → B) Codex CLI → C) Cursor → D) OpenCode → E) Várias — escolho mais do que uma

Then generate INIT.md (see Template section below) and show the review gate.


Path B — No ROADMAP (Full Interview)

Ask these 5 questions, one at a time. Use plain language — assume the user is non-technical.

Q1 (free text):

"O que queres construir? Descreve com as tuas palavras."

Q2 (multiple choice):

"Já existe código ou começas do zero?" → A) Começo do zero → B) Já existe código → C) Não tenho a certeza

If answer is B: run stack inference silently (see Stack Inference section).

Q3 (multiple choice):

"Que tipo de projecto é?" → A) App web (abre no browser) → B) App móvel (iPhone ou Android) → C) API ou serviço de backend → D) Análise de dados ou automação → E) Outro

Q4 (multiple choice):

"Em que dispositivo ou plataforma deve correr?" → A) Browser (qualquer dispositivo) → B) iPhone / iPad → C) Android → D) Desktop (Windows / Mac / Linux) → E) Servidor / cloud

Q5 (multiple choice):

"Que ferramentas de IA estás a usar para escrever código?" → A) Claude Code → B) Codex CLI → C) Cursor → D) OpenCode → E) Várias — escolho mais do que uma

After Q5: generate INIT.md, show review gate, then offer to create ROADMAP.md:

"Queres que eu crie um ROADMAP.md para este projecto com base no que descreveste? É útil para futuras sessões."


Stack Inference (Path B, answer B to Q2 only)

Run silently after the user answers B. Do NOT run for new projects (answer A or C).

git ls-files | sed 's/.*\.//' | sort | uniq -c | sort -rn | head -10

Extension → stack mapping:

  • .kt → Kotlin / Android (agents: kotlin-reviewer, compose-ui)
  • .swift → Swift / iOS (agents: swift-reviewer)
  • .py → Python (agents: python-reviewer)
  • .ts or .tsx → TypeScript / React (agents: code-reviewer)
  • .go → Go (agents: go-reviewer)
  • .rs → Rust (agents: rust-reviewer)
  • .dart → Flutter (agents: flutter-reviewer)
  • .java → Java / Android (agents: kotlin-reviewer as fallback)

Confidence threshold: if one extension accounts for >40% of tracked files, pre-fill Q3/Q4 and confirm with user. If ambiguous, ask normally.


Review Gate

After generating INIT.md, display ONLY the ## O que entendi section and ask:

"Está correcto? Falta alguma coisa?"

  • User says "ok" / "sim" / "yes" → run /orchestrate init automatically
  • User describes a correction → update the relevant INIT.md section, show ## O que entendi again
  • After 3 correction rounds without approval → ask user to edit INIT.md manually: "Não consegui perceber a correcção. Por favor edita o INIT.md directamente e diz 'ok' quando estiver pronto."

Codex Warning

Show this before the final approval:

"Última coisa: o Codex vai pedir para aprovares um script de segurança na próxima sessão. Clica em 'Trust' para continuar — é o script de estado da A Team."


INIT.md Template

Generate this file at the project root. Fill each field from the interview answers or ROADMAP extraction.

## O que entendi
[Plain language summary: what the project is, inferred stack, active AI platforms, agent count after init]

Se algo estiver errado, edita este ficheiro antes de continuar.

---

# INIT.md — [Project Name]

> Run `/orchestrate init` after reviewing this file.

## Project Overview

**Name:** [from Q1 or ROADMAP H1]
**Type:** [from Q3: web app / mobile app / API / data / other]
**Status:** [New project / Active development]
**Description:** [from Q1 free text or ROADMAP description]

## Languages & Stack

[Check all that apply — inferred from ROADMAP or stack scan]
- [ ] Kotlin
- [ ] Swift
- [ ] Python
- [ ] TypeScript / JavaScript
- [ ] Go
- [ ] Rust
- [ ] Flutter / Dart
- [ ] Other: ___

**Framework / UI:** [inferred or left blank]
**Database:** [inferred or left blank]
**Build system:** [inferred or left blank]

## Compliance Scope

[Extracted from ROADMAP compliance mentions, or left unchecked]
- [ ] GDPR
- [ ] Child privacy / COPPA
- [ ] Local-first / no external data
- [ ] HIPAA
- [ ] PCI-DSS

## Active AI Platforms

[From Q5 or Path A question]
- [ ] Claude Code
- [ ] Codex CLI
- [ ] Cursor
- [ ] OpenCode

## Non-Negotiable Rules

[Extracted from ROADMAP "princípios não-negociáveis" section, or left blank]

## Active Work / Next Steps

[Extracted from ROADMAP "Próximo" / "Next" section, or left blank]

## Agents to Prune

[Inferred from stack — list agents irrelevant to this project's languages/domain]

04 同仓库更多 Skills

hashgraph-online
hashgraph-online/awesome-codex-plugins

migrate-to-e2e

Moves an existing Playwright, Cypress, Selenium, Appium, Detox, Maestro, or TesterArmy suite to e2e one slice at a time, with the old suite running beside the new one until each ported test is proven. Keeps a MIGRATION.md progress file so any later session resumes where the last stopped. Optionally plants a defect to show both suites catch it. Use for /migrate-to-e2e, "migrate our tests to e2e", "port these Maestro flows", "move our TesterArmy tests", or when a project has an e2e config and an older suite still running.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

e2e

Writes, runs, and debugs end-to-end tests for web and mobile apps with e2e, the agentic test runner, which mixes agent.act goals with exact locator assertions on a Playwright browser or an iOS or Android device. Also drives a project's running app live over the e2e MCP server to look at a screen, check a UI change, take a screenshot or record a video, and try a locator; and runs bug bashes, parallel e2e explore runs with each bug proven by a repro test. Use when a project depends on e2e or has an e2e.config.ts, when asked for end-to-end, browser, mobile, or agentic UI tests, to check a change in the running app, to QA or bug bash an app or branch, or when an e2e run fails.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

create-verification-skill

Generate a project-local verify-<app> skill on top of e2e, so any coding agent can launch the app, check it is worth driving, exercise a feature the way a user does, keep evidence, and bug bash it. Writes the skill, a feature map with one explore charter per feature, and proves it once. Use for /create-verification-skill, "make a verification skill for this repo", or when a project has no scripted way to prove UI behavior.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

deja-search

deja-vu memory — search the user's past AI coding sessions with the deja CLI. Holds what they already settled and never wrote into the repo (naming rules, banned or already-tried tools, required checks before merge, preferred values, who owns what). Use before your first edit in a task, before you call a change done or ready to merge, before you recommend a tool, library, design or reviewer, before re-debugging an error that may already be solved, when they say 'didn't we fix this before', 'what did we decide about X' or 'I already have that', and before telling them something on this machine does not exist.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

deja-reflect

End-of-session reflection and memory curation with deja-vu. Review what was decided, what commands broke and what repaired them, and record durable project rules or decisions with deja remember so the next session starts with them.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

revid

Make, edit, export and post AI videos with Revid (revid.ai) through the Revid MCP server. Use whenever the user wants a short-form or long-form video made (from an idea, a script, a URL, an article, a PDF, a tweet, a product page, their own clips, or a song), wants to find what is going viral on TikTok, YouTube, Instagram or X before making something, wants to edit or re-export an existing Revid project, generate a standalone AI image or clip, make a series episode, or publish and schedule a finished video to connected social accounts. Covers setup (hosted MCP with OAuth, or an API key), the 36 tools, credits, polling and the rules for publishing. Routes to the more specific revid-* skills when one fits.

图像与多媒体Skill与智能体开发
★ 1.6k