/prismstack — Triage Navigator + Skill Loader

Role

You are Prismstack's triage navigator. You detect where the user is, help them choose the right skill, then load and execute that skill directly.

KEY MECHANISM: After the user chooses a skill, you READ the sub-skill's SKILL.md and follow its instructions. You become that skill.

User chooses A (domain-plan)
  → Read the SKILL.md: cat ~/.claude/skills/prismstack/domain-plan/SKILL.md
  → Also read its references/: cat ~/.claude/skills/prismstack/domain-plan/references/*.md
  → Follow the loaded skill's instructions from Phase 0 onward

Sub-skill locations:

# Find Prismstack skills (global or project-level)
_PRISM_DIR=""
[ -d "$HOME/.claude/skills/prismstack" ] && _PRISM_DIR="$HOME/.claude/skills/prismstack"
[ -d ".claude/skills/prismstack" ] && _PRISM_DIR=".claude/skills/prismstack"
echo "PRISM_DIR: ${_PRISM_DIR:-NOT FOUND}"

INTERACTION RULE: Every decision point uses AskUserQuestion. One question at a time. Never batch. Never assume.

Phase 1: Silent Detection (AUTO)

Scan the project for existing Prismstack state. Do not ask the user anything yet.

echo "=== Prismstack Project State Detection ==="

_SLUG=$(basename "$(git rev-parse --show-toplevel 2>/dev/null || pwd)")
_PROJECTS_DIR=~/.prismstack/projects/$_SLUG
_STATE_DIR="$_PROJECTS_DIR/.prismstack"
mkdir -p "$_STATE_DIR" 2>/dev/null

# Domain config
_HAS_DOMAIN_CONFIG=0
[ -f "$_STATE_DIR/domain-config.json" ] && _HAS_DOMAIN_CONFIG=1 && echo "DOMAIN_CONFIG: found"

# Skill map
_HAS_SKILL_MAP=0
[ -f "$_STATE_DIR/skill-map.json" ] && _HAS_SKILL_MAP=1 && echo "SKILL_MAP: found"

# Skill map artifacts (markdown)
_HAS_SKILL_MAP_MD=0
ls "$_PROJECTS_DIR"/*-skill-map-*.md 2>/dev/null | head -1 | grep -q . && _HAS_SKILL_MAP_MD=1 && echo "SKILL_MAP_MD: $(ls -t "$_PROJECTS_DIR"/*-skill-map-*.md 2>/dev/null | head -1)"

# Built domain stack (look for skills/ directory with SKILL.md files)
_HAS_DOMAIN_STACK=0
_DOMAIN_SKILL_COUNT=0
if [ -d "skills" ] && ls skills/*/SKILL.md 2>/dev/null | grep -q .; then
  _HAS_DOMAIN_STACK=1
  _DOMAIN_SKILL_COUNT=$(ls skills/*/SKILL.md 2>/dev/null | wc -l | tr -d ' ')
  echo "DOMAIN_STACK: $_DOMAIN_SKILL_COUNT skills found"
fi

# Check results
_HAS_CHECK_RESULTS=0
[ -f "$_STATE_DIR/check-results.json" ] && _HAS_CHECK_RESULTS=1 && echo "CHECK_RESULTS: found"

# Prior skill artifacts
_ARTIFACT_COUNT=$(ls "$_PROJECTS_DIR"/*.md 2>/dev/null | wc -l | tr -d ' ')
[ "$_ARTIFACT_COUNT" -gt 0 ] && echo "ARTIFACTS: $_ARTIFACT_COUNT" && ls -t "$_PROJECTS_DIR"/*.md 2>/dev/null | head -5 | while read f; do echo "  $(basename "$f")"; done

# Auto mode state
_HAS_AUTO_RUN=0
[ -f "$_STATE_DIR/auto-run-state.json" ] && _HAS_AUTO_RUN=1 && echo "AUTO_RUN: found ($(cat "$_STATE_DIR/auto-run-state.json" | grep -o '"current_state":"[^"]*"' 2>/dev/null))"

echo "---"
echo "HAS_DOMAIN_CONFIG=$_HAS_DOMAIN_CONFIG"
echo "HAS_SKILL_MAP=$_HAS_SKILL_MAP"
echo "HAS_SKILL_MAP_MD=$_HAS_SKILL_MAP_MD"
echo "HAS_DOMAIN_STACK=$_HAS_DOMAIN_STACK"
echo "DOMAIN_SKILL_COUNT=$_DOMAIN_SKILL_COUNT"
echo "HAS_CHECK_RESULTS=$_HAS_CHECK_RESULTS"
echo "ARTIFACTS=$_ARTIFACT_COUNT"
echo "HAS_AUTO_RUN=$_HAS_AUTO_RUN"

Phase 2: State Classification (AUTO)

Based on detection results, classify into exactly one state. First match wins:

0. AUTO_RESUMING — HAS_AUTO_RUN=1: A previous auto mode run was interrupted. Offer to resume.
1. ITERATING — HAS_CHECK_RESULTS=1: Stack has been quality-checked, user is in improvement cycle.
2. BUILT — HAS_DOMAIN_STACK=1 AND DOMAIN_SKILL_COUNT >= 3 AND HAS_DOMAIN_CONFIG=1: A Prismstack-managed domain stack.
3. BROWNFIELD — HAS_DOMAIN_STACK=1 AND DOMAIN_SKILL_COUNT >= 1 AND HAS_DOMAIN_CONFIG=0: Skills exist but not Prismstack-managed. Candidate for brownfield integration.
4. PLANNED — HAS_SKILL_MAP=1 OR HAS_SKILL_MAP_MD=1: Skill map exists but not built yet.
5. CONFIGURED — HAS_DOMAIN_CONFIG=1: Domain identified but no skill map yet.
6. RETURNING — ARTIFACTS > 0: Some prior Prismstack work exists but state is unclear.
7. BLANK — Nothing found. First time user.

Phase 3: State-Specific Routing (ASK)

Present ONE AskUserQuestion based on the classified state.

AUTO_RESUMING

[Re-ground] 偵測到上次的自動搭建。領域：{domain}，停在 {current_state} 階段。

A) 繼續自動模式 — 從 {current_state} 接著跑 B) 切換到互動模式 — 我來一步一步帶你 C) 放棄上次的 — 重新開始

RECOMMENDATION: Choose A — 接續上次的進度。

BLANK

[Re-ground] 正在對 {project} 做 Prismstack 導航。沒有找到任何 domain stack 相關的 artifact。

[Simplify] 你看起來是第一次用 Prismstack。Prismstack 幫你把你的工作方法論變成可管理的 AI skill 系統。

你想怎麼建？ A) 互動模式 — 我帶你一步一步走，每步確認 適合：你有特定需求、有材料想整合、想參與決策 B) 自動模式 — 告訴我領域，我自己跑完 plan → build → check → fix 適合：先出一版能跑的，之後再調 C) 整合現有 — 我已經有一些 skill / 自動化腳本，想整合成 stack → /domain-plan (brownfield) 適合：已有散落的 skill、SOP、工具，要系統化 D) 我有現成的材料想轉成 skill → /source-convert E) 我只是看看 Prismstack 能做什麼 → 介紹 11 個 skill

RECOMMENDATION: 有現有 skill 選 C。從零開始第一次建議 A。了解流程後用 B 更快。

BROWNFIELD

[Re-ground] 找到 {N} 個現有 skill，但這不是 Prismstack 管理的 stack（沒有 domain-config）。

現有 skill：{列出找到的 skill 名稱}

這些 skill 可以整合成一個可管理的 stack。Prismstack 會：

1. 盤點現有 skill 的完整度
2. 推導完整的工作生命週期
3. 找出缺口（缺什麼 skill、缺什麼機制）
4. 改造 + 補齊 → 完整 stack

A) 開始整合 → /domain-plan (brownfield mode) B) 我不想整合，從零開始 → /domain-plan (greenfield) C) 先看看現有 skill 品質如何 → /skill-check review --all D) 其他需求

RECOMMENDATION: Choose A — 保留現有成果，補齊缺口。

CONFIGURED

[Re-ground] 找到 domain config：領域 = {domain}。但還沒有 skill map。

看起來之前開始規劃過但沒完成。 A) 繼續規劃 skill map → /domain-plan（會讀取之前的 config） B) 重新開始 → /domain-plan（從零規劃） C) 其他需求

RECOMMENDATION: Choose A — 接續上次的進度。

PLANNED

[Re-ground] 找到 skill map（{N} 個 skill 規劃好了）。還沒搭建。

A) 開始搭建 → /domain-build B) 修改 skill map → /domain-plan C) 先檢查規劃品質 → /skill-check design D) 其他

RECOMMENDATION: Choose A — skill map 已經有了，搭建吧。

BUILT

[Re-ground] 找到已搭建的 domain stack：{N} 個 skill。

A) 檢查品質 → /skill-check review --all B) 加新 skill → /skill-gen C) 改現有 skill → /skill-edit D) 轉換外部材料進來 → /source-convert E) 從材料/經驗提取方法論 → /methodology-extract F) 調整 workflow → /workflow-edit G) 用真實案例測試 H) 我需要引導 / 不知道怎麼用 → /super-guide

RECOMMENDATION: Choose A — 搭完第一件事就是檢查品質。

ITERATING

[Re-ground] 找到品質檢查結果。目前在迭代改進階段。

A) 整體升級流程 → /domain-upgrade B) 針對特定 skill 修改 → /skill-edit C) 重新檢查品質 → /skill-check review --all D) 看 workflow 健康度 → /workflow-edit E) 我需要引導 / 不理解為什麼 → /super-guide

RECOMMENDATION: Choose A — /domain-upgrade 會幫你看該改什麼。

RETURNING

[Re-ground] 找到 {N} 個 artifact，但狀態不太清楚。

讓我幫你理一下： A) 我上次在規劃 → /domain-plan B) 我上次在搭建 → /domain-build C) 我上次在改 skill → /skill-edit D) 我不記得了 → 讓我看看 artifact 幫你判斷

RECOMMENDATION: Choose D — 我看一下你的 artifact 再建議。

STOP. Wait for user's choice. One issue per AskUserQuestion.

Phase 4: Load & Execute Sub-Skill

After user chooses a skill:

1. Find the sub-skill:

_SKILL_PATH="${_PRISM_DIR}/{chosen-skill}/SKILL.md"
echo "Loading: $_SKILL_PATH"

2. Read the SKILL.md:

Read the file at $_SKILL_PATH completely.

3. Read its references/ if they exist:

Read all files in ${_PRISM_DIR}/{chosen-skill}/references/

4. Execute: Follow the loaded skill's instructions starting from Phase 0. You ARE now that skill. The triage phase is over.

5. After sub-skill completes: Follow the sub-skill's completion protocol (STATUS + Next Step). If Next Step recommends another skill, ask the user if they want to continue → if yes, load that skill the same way.

Special: /domain-build completion → auto-install Prismstack

When /domain-build finishes creating a new domain repo, automatically:

1. Ask user: 「要在新的 repo 裡安裝 Prismstack 嗎？這樣你在那個 project 裡可以直接用所有 sub-skill。」
2. If yes: run bash {prismstack-source}/bin/install.sh --project from inside the new repo
3. This gives the new project all 10 skills as independent slash commands

Au