ultimate-pi 0.1.2 → 0.1.3

package/.pi/prompts/harness-setup.md

@@ -0,0 +1,509 @@
+---
+description: Full harness bootstrap — obsidian wiki scaffold, CLI tools install, pi extension packages, and verification. Run once per project.
+argument-hint: "[--skip-wiki] [--skip-tools] [--force] [--wiki-path <path>]"
+---
+
+# harness-setup — Full Harness Bootstrap
+
+Bootstraps the complete ultimate-pi agentic harness: Obsidian wiki vault, CLI tools, pi extension packages, configuration files, and verification. Idempotent — safe to re-run, skips what's already installed.
+
+## Step 0 — Pre-flight Environment Check
+
+```bash
+which node && node --version
+which npm && npm --version
+which git && git --version
+```
+
+Block if node < 18, npm < 9, or git missing. Report versions and continue.
+
+Read `.pi/auto-commit.json` for co-author + branch config. Read `.pi/settings.json` for extension packages list and `wiki_path` field (if present).
+
+## Step 0.5 — Resolve Wiki Vault Path + Detect Existing Vault
+
+Resolve the wiki vault path with this priority:
+
+1. **CLI flag** `--wiki-path <path>` (highest priority — user explicitly passed it)
+2. **Env var** `VAULT_WIKI_PATH` (set by project or user shell config)
+3. **Settings file** `.pi/settings.json` → `wiki_path` field
+4. **Default** `vault/wiki` (relative to project root)
+
+```bash
+# Resolve logic
+if [ -n "$CLI_WIKI_PATH" ]; then
+  WIKI_PATH="$CLI_WIKI_PATH"
+elif [ -n "$VAULT_WIKI_PATH" ]; then
+  WIKI_PATH="$VAULT_WIKI_PATH"
+else
+  WIKI_PATH="$(node -e "try{const s=require('./.pi/settings.json');if(s.wiki_path)process.stdout.write(s.wiki_path)}catch(e){}" 2>/dev/null)"
+  WIKI_PATH="${WIKI_PATH:-vault/wiki}"
+fi
+
+# Normalize to absolute path (relative paths are relative to project root)
+if [[ "$WIKI_PATH" != /* ]]; then
+  WIKI_PATH="$(pwd)/$WIKI_PATH"
+fi
+
+# Detect if a vault already exists at the resolved path
+# Vault indicators: .obsidian/ folder OR wiki/index.md OR wiki/log.md
+if [ -d "$WIKI_PATH/.obsidian" ] || [ -f "$WIKI_PATH/index.md" ] || [ -f "$WIKI_PATH/log.md" ]; then
+  VAULT_EXISTS=true
+  # Quick stats
+  PAGE_COUNT=$(find "$WIKI_PATH" -name "*.md" -type f 2>/dev/null | wc -l)
+  LAST_LOG=$(head -5 "$WIKI_PATH/log.md" 2>/dev/null || echo "unknown")
+else
+  VAULT_EXISTS=false
+fi
+```
+
+**Present to user explicitly — context-aware:**
+
+### Case A: Env var was set + vault found
+
+> "`VAULT_WIKI_PATH` is set to `$VAULT_WIKI_PATH`. Vault detected: $PAGE_COUNT pages, last updated $LAST_LOG. Confirm this is the correct vault?"
+
+### Case B: Env var was set + NO vault found
+
+> "`VAULT_WIKI_PATH` is set to `$VAULT_WIKI_PATH`, but no vault found there (no `.obsidian/`, `index.md`, or `log.md`). Where should we create the wiki vault? (Press Enter to use this path, or type a different path)"
+
+### Case C: Settings had wiki_path + vault found
+
+> "Found `wiki_path: <path>` in `.pi/settings.json`. Vault detected: $PAGE_COUNT pages, last updated $LAST_LOG. Use this vault?"
+
+### Case D: No env var, no settings → defaulting
+
+> "No `VAULT_WIKI_PATH` env var or `wiki_path` in settings found. Default path: `$(pwd)/vault/wiki`. Where should we create the wiki vault? (Press Enter for default, or type a different path)"
+
+### Case E: All sources agree on path + vault found
+
+> "Wiki vault: `$WIKI_PATH` ($PAGE_COUNT pages, last updated $LAST_LOG). Use this?"
+
+**Wait for user confirmation before proceeding.** This is the ONLY blocking question. If user provides a different path, update `WIKI_PATH`, re-detect, and re-confirm.
+
+**DX principles:**
+- Always show the resolved absolute path so user knows exactly where files go.
+- Always show the *source* of the path (env var, settings, default) so user understands WHY this path was chosen.
+- When no vault is detected, ask explicitly — don't just silently create at the default.
+- Don't check for `.obsidian/` in project root as a fallback. The vault can live anywhere.
+
+## Step 1 — Wiki Scaffold
+
+Run the `/wiki` prompt flow using the user-confirmed `$WIKI_PATH`:
+
+1. If `VAULT_EXISTS=true` (detected in Step 0.5): skip scaffold. Move to Step 2.
+2. If `VAULT_EXISTS=false`: ask ONE question: "What is this vault for?"
+3. Scaffold full wiki structure based on answer at `$WIKI_PATH`:
+   ```
+   $WIKI_PATH/
+   ├── index.md          # master catalog
+   ├── log.md            # chronological operations log
+   ├── hot.md            # recent context summary
+   ├── overview.md       # executive summary
+   ├── sources/          # source document summaries
+   ├── entities/         # people, orgs, products, repos
+   ├── concepts/         # ideas, patterns, frameworks
+   ├── decisions/        # ADRs + design decisions
+   ├── questions/        # filed research answers
+   ├── modules/          # code module documentation
+   ├── flows/            # pipeline/process flows
+   ├── consensus/        # debate verdict records
+   ├── meta/             # dashboards, lint reports
+   └── components/       # reusable sub-modules
+   ```
+4. Create `$WIKI_PATH/../.vault-meta/` with vault metadata (at project root level, not inside wiki).
+5. Create vault `AGENTS.md` in project root with mode, purpose, conventions, operations.
+6. Initialize wiki git tracking if not already present.
+7. Write initial `$WIKI_PATH/hot.md` with setup timestamp.
+8. **Save resolved path** to `.pi/settings.json` (merge `"wiki_path": "<relative-path>"` into settings) for future sessions.
+
+## Step 2 — Install Global CLI Packages
+
+Check each package first. Install only if missing unless `--force` flag.
+
+### 2.1 — firecrawl-cli (Web Search + Scrape + Crawl + Interact + Download + Parse)
+
+```bash
+if ! command -v firecrawl &>/dev/null || [ "$FORCE" = "true" ]; then
+	npm install -g firecrawl-cli@latest
+fi
+```
+
+Verify:
+```bash
+firecrawl --status
+```
+
+If not authenticated, offer browser login or API key:
+```bash
+firecrawl login --browser
+# OR
+firecrawl login --api-key "<key>"
+```
+
+Install skills and run quick smoke test:
+```bash
+firecrawl setup skills
+mkdir -p .firecrawl
+firecrawl scrape "https://firecrawl.dev" -o .firecrawl/install-check.md
+```
+
+### 2.2 — defuddle-cli (Web Page Cleaner)
+
+```bash
+if ! command -v defuddle &>/dev/null || [ "$FORCE" = "true" ]; then
+	npm install -g defuddle-cli
+fi
+```
+
+Verify: `defuddle --version`
+
+### 2.3 — ctx7 (Context7 Library Docs + Skills Management)
+
+```bash
+if ! command -v ctx7 &>/dev/null || [ "$FORCE" = "true" ]; then
+	npm install -g ctx7@latest
+fi
+```
+
+Verify: `ctx7 --help`
+
+Offer login for higher rate limits:
+```bash
+ctx7 login
+ctx7 whoami
+```
+
+### 2.4 — agent-browser (Vercel Labs Browser Automation for AI Agents)
+
+```bash
+if ! command -v agent-browser &>/dev/null || [ "$FORCE" = "true" ]; then
+	npm install -g agent-browser
+fi
+```
+
+Verify:
+```bash
+agent-browser --version
+```
+
+Create config directory:
+```bash
+mkdir -p .pi/harness
+```
+
+Create default browser config if missing:
+```bash
+if [ ! -f .pi/harness/browser.json ]; then
+	echo '{"headless": true, "timeout": 30000, "viewport": {"width": 1280, "height": 720}}' > .pi/harness/browser.json
+fi
+```
+
+### 2.5 — ck-search (Semantic Code Search)
+
+```bash
+if ! command -v ck &>/dev/null || [ "$FORCE" = "true" ]; then
+	npm install -g @beaconbay/ck-search
+fi
+```
+
+Verify: `ck --version`
+
+Register as MCP server (if Claude MCP available):
+```bash
+claude mcp list 2>/dev/null && claude mcp add ck-search -s user -- ck --serve || echo "MCP not available — ck will be used as CLI only"
+```
+
+### 2.6 — fallow (Codebase Intelligence: Dead Code, Duplication, Complexity)
+
+```bash
+if ! command -v fallow &>/dev/null || [ "$FORCE" = "true" ]; then
+	npm install -g fallow
+fi
+```
+
+Verify:
+```bash
+fallow --version
+```
+
+Set up baseline for existing codebases:
+```bash
+fallow audit --gate all --format json > .fallow-baseline.json 2>/dev/null || echo "No baseline yet — will be created on first gate run"
+```
+
+### 2.7 — biome (Lint + Format Gate)
+
+```bash
+if ! command -v biome &>/dev/null || [ "$FORCE" = "true" ]; then
+	npm install -g @biomejs/biome
+fi
+```
+
+Check if project already has biome config:
+```bash
+ls biome.json 2>/dev/null && echo "biome.json found — using project config" || echo "No biome.json — using defaults"
+```
+
+Verify: `biome --version`
+
+### 2.8 — gh CLI (GitHub Issues Spec Storage — ADR-025)
+
+```bash
+if ! command -v gh &>/dev/null || [ "$FORCE" = "true" ]; then
+	echo "gh CLI not found. Install: https://cli.github.com/"
+fi
+```
+
+Verify and authenticate:
+```bash
+gh auth status && echo "gh authenticated" || echo "Run: gh auth login"
+```
+
+Create harness labels if authenticated:
+```bash
+if gh auth status &>/dev/null; then
+	gh label create "harness" --color "0366d6" --description "Agentic harness managed" 2>/dev/null
+	gh label create "harness-spec" --color "0e8a16" --description "Hardened specification" 2>/dev/null
+	gh label create "harness-plan" --color "fbca04" --description "Structured plan generated" 2>/dev/null
+	gh label create "harness-critic" --color "d73a4a" --description "Adversarial review" 2>/dev/null
+fi
+```
+
+## Step 3 — Pi Extension Packages
+
+Install pi extension packages from `.pi/settings.json`:
+
+```bash
+cd .pi/npm
+npm install
+```
+
+Verify each package:
+
+| Package | Purpose | Phase |
+|---------|---------|-------|
+| `@posthog/pi` | Analytics event capture | F0 |
+| `pi-lean-ctx` | Context runtime (read/bash/find/grep/MCP bridge) | F0 |
+| `@tintinweb/pi-subagents` | L4 critic sub-agent spawn/control | P16 |
+| `@yeliu84/pi-model-router` | Per-turn intelligent model routing (auto high/medium/low tier selection) | F0 |
+
+## Step 3.5 — Model Router Configuration
+
+Verify the model router is installed and configured:
+
+```bash
+# Verify package installed
+ls .pi/npm/node_modules/@yeliu84/pi-model-router/package.json 2>/dev/null && echo "✓ model-router package" || echo "✗ model-router package — run: cd .pi/npm && npm install"
+
+# Verify config exists
+ls .pi/model-router.json 2>/dev/null && echo "✓ model-router.json" || echo "✗ model-router.json — see .pi/npm/node_modules/@yeliu84/pi-model-router/model-router.example.json"
+```
+
+Do NOT block. If config exists, continue. If missing, warn in report and continue.
+
+**Router activation happens automatically** — the agent should output the following as its next message (this activates the router in the current session):
+
+> `/router profile auto`
+
+The pi TUI will intercept this and activate the `auto` profile. Then continue to Step 4.
+
+## Step 4 — Configuration Files
+
+### 4.1 — .gitignore Entries
+
+Ensure `.gitignore` contains:
+```
+.firecrawl/
+.raw/
+.vault-meta/
+.fallow-baseline.json
+.pi/harness/critics/
+.pi/harness/plans/
+.pi/harness/specs/
+```
+
+### 4.2 — Vault AGENTS.md
+
+If not already created by Step 1, create a minimal `AGENTS.md` in project root:
+
+```markdown
+# ultimate-pi: Agentic Harness Wiki
+
+Mode: B (Engineering Research + System Design)
+Purpose: Knowledge base for the ultimate-pi agentic coding harness — architecture, research, decisions, implementation plans.
+Owner: pi-mono + user
+Created: $(date +%Y-%m-%d)
+Wiki vault path: `$WIKI_PATH`
+
+## Structure
+
+`$WIKI_PATH/`
+├── index.md → master catalog
+├── log.md → chronological operations
+├── hot.md → recent context cache
+├── concepts/ → ideas, patterns, frameworks (harness layers, drift detection, skill-first architecture)
+├── decisions/ → ADRs for every architectural choice
+├── modules/ → code module docs (harness implementation plan, pipeline)
+├── sources/ → reference materials
+├── entities/ → tools, platforms, people
+├── questions/ → filed research answers
+├── consensus/ → debate verdicts
+└── flows/ → pipeline diagrams
+
+## Conventions
+
+- YAML frontmatter required: type, status, created, updated, tags
+- Wikilinks: [[Page Name]] format
+- .raw/ is immutable source storage
+- $WIKI_PATH/index.md updated on every ingest
+- $WIKI_PATH/log.md is append-only, newest at top
+
+## Cross-Project Reference
+
+Other projects can reference this vault:
+1. Read $WIKI_PATH/hot.md (~500 tokens)
+2. Read $WIKI_PATH/index.md if needed
+3. Drill into specific topics as needed
+
+## Path Resolution
+
+This vault's filesystem path is set via:
+- Env var: `VAULT_WIKI_PATH` (current: `$VAULT_WIKI_PATH` or unset)
+- Settings: `.pi/settings.json` → `wiki_path`
+- Default fallback: `vault/wiki`
+```
+
+## Step 5 — Verification
+
+Run full verification suite using resolved `$WIKI_PATH`:
+
+```bash
+# CLI tools
+firecrawl --status 2>/dev/null && echo "✓ firecrawl" || echo "✗ firecrawl"
+defuddle --version 2>/dev/null && echo "✓ defuddle" || echo "✗ defuddle"
+ctx7 --help 2>/dev/null && echo "✓ ctx7" || echo "✗ ctx7"
+agent-browser --version 2>/dev/null && echo "✓ agent-browser" || echo "✗ agent-browser"
+ck --version 2>/dev/null && echo "✓ ck-search" || echo "✗ ck-search"
+fallow --version 2>/dev/null && echo "✓ fallow" || echo "✗ fallow"
+biome --version 2>/dev/null && echo "✓ biome" || echo "✗ biome"
+gh --version 2>/dev/null && echo "✓ gh" || echo "✗ gh"
+
+# pi extensions
+cd .pi/npm && npm ls 2>/dev/null && echo "✓ pi extensions" || echo "✗ pi extensions"
+
+# wiki vault (using resolved WIKI_PATH)
+ls "$WIKI_PATH/index.md" 2>/dev/null && echo "✓ wiki vault" || echo "✗ wiki vault"
+ls "$WIKI_PATH/hot.md" 2>/dev/null && echo "✓ wiki hot cache" || echo "✗ wiki hot cache"
+
+# model router
+ls .pi/npm/node_modules/@yeliu84/pi-model-router/package.json 2>/dev/null && echo "✓ model-router package" || echo "✗ model-router package"
+ls .pi/model-router.json 2>/dev/null && echo "✓ model-router config" || echo "✗ model-router config"
+
+# settings persistence
+grep -q 'wiki_path' .pi/settings.json 2>/dev/null && echo "✓ wiki path saved to settings" || echo "! wiki path not saved"
+
+# gitignore entries
+grep -q '.firecrawl/' .gitignore 2>/dev/null && echo "✓ .gitignore" || echo "! .gitignore missing entries"
+```
+
+## Step 6 — Wiki Hot Cache Bootstrap
+
+Write initial `$WIKI_PATH/hot.md`:
+```markdown
+---
+type: meta
+title: "Hot Cache"
+updated: $(date -u +%Y-%m-%dT%H:%M:%S)
+created: $(date +%Y-%m-%d)
+tags: []
+status: active
+---
+
+# Recent Context
+
+## Last Updated
+$(date +%Y-%m-%d). Harness setup completed. All CLI tools installed, wiki vault scaffolded at `$WIKI_PATH`, pi extensions configured.
+
+## Key Facts
+- Harness runs 8-layer mandatory pipeline (L1 → L2 → L2.5 → L3 → L4 → P20 → L5 → L6 → L7 → L8)
+- Implementation: skill-first v2 — 4 code files, 6 harness skills
+- Wiki vault: Mode B (Engineering Research + System Design)
+- Wiki path: `$WIKI_PATH`
+- ADRs filed in `$WIKI_PATH/decisions/`
+- Consensus debate verdicts filed in `$WIKI_PATH/consensus/`
+
+## Recent Changes
+- Created: Full wiki vault structure
+
+- Installed: firecrawl-cli, defuddle-cli, ctx7, agent-browser, ck-search, fallow, biome, gh
+
+## Active Threads
+- Next: run first spec through pipeline with `/harness "task description"`
+```
+
+## Step 7 — Report
+
+Output summary table:
+
+| Component | Status | Detail |
+|-----------|--------|--------|
+| Wiki Vault | ✓/✗ | Path: `$WIKI_PATH` — scaffold complete / task pending |
+| firecrawl-cli | ✓/✗ | Auth: yes/no |
+| defuddle-cli | ✓/✗ | Version |
+| ctx7 | ✓/✗ | Login: yes/no |
+| agent-browser | ✓/✗ | Config: .pi/harness/browser.json |
+| ck-search | ✓/✗ | MCP: registered/CLI-only |
+| fallow | ✓/✗ | Baseline: created/pending |
+| biome | ✓/✗ | Project config: found/default |
+| gh CLI | ✓/✗ | Auth: yes/no |
+| pi extensions | ✓/✗ | 4 packages |
+| model router | ✓/✗ | Package + config verified, activation via `/router profile auto` |
+
+| .gitignore | ✓/✗ | 6 entries added |
+| wiki_path in settings | ✓/✗ | Persisted to .pi/settings.json |
+
+Next steps:
+1. If tools missing: re-run with `--force` or install individually
+2. If wiki not scaffolded: run `/wiki` prompt
+3. If gh not authenticated: `gh auth login`
+4. First harness run: `/harness "your task description"`
+5. To change wiki path later: update `VAULT_WIKI_PATH` env var or `.pi/settings.json` → `wiki_path` field
+
+## Guard Rails
+
+- **Internet required**: Several tools need npm registry access. Block if offline.
+- **Wiki path must be writable**: Check `test -w "$(dirname "$WIKI_PATH")"` before scaffold. Block if not writable.
+- **Wiki path outside project**: Allowed (e.g., `~/vaults/my-project`). Cross-project vault sharing is supported.
+- **Node.js >= 18 required**: Some pi packages use modern Node APIs.
+- **Idempotent**: All checks skip if already installed. `--force` overrides.
+- **No destructive actions**: Creates files only if missing. Never overwrites existing wiki content.
+- **Wiki safety**: Scaffold only creates structure, never modifies existing wiki content.
+- **Partial success**: If some tools fail, report which and continue. User can fix individually.
+- **Rate limits**: ctx7 login is optional. firecrawl auth is required for web operations.
+- **Settings persistence**: The resolved `wiki_path` is saved to `.pi/settings.json` so future sessions auto-detect it.
+
+## Error Handling
+
+| Error | Action |
+|-------|--------|
+| Node < 18 | Block. Report required version. |
+| npm not found | Block. Suggest install method per OS. |
+| Wiki path not writable | Block. Show path and suggest `chmod` or different path. |
+| Wiki path already has vault + `--skip-wiki` not set | Report state, ask if user wants to skip or force re-scaffold. |
+| `VAULT_WIKI_PATH` and `wiki_path` in settings disagree | Resolve with priority (env var wins), warn user about mismatch, ask to confirm. |
+| firecrawl auth failed | Show manual login instructions. Continue with other tools. |
+| gh not installed | Show GitHub CLI install link. Skip label creation. |
+| pi packages install fail | Show error output. Check npm permissions. |
+| wiki already exists | Report state, skip scaffold, continue other steps. |
+| biome.json missing | Create minimal config. |
+| settings.json not writable | Warn. Wiki path won't persist across sessions. |
+| No internet | Block for tool installs. Continue for wiki-only steps if `--skip-tools`. |
+
+## Flags
+
+| Flag | Effect |
+|------|--------|
+| `--skip-wiki` | Skip Step 1 (wiki scaffold). Use when wiki already exists. |
+| `--skip-tools` | Skip Step 2 (CLI tool installs). Use when tools already set up. |
+| `--force` | Reinstall all tools even if already present. Overwrite existing files. |
+| `--wiki-path <path>` | Override wiki vault path. Absolute or relative to project root. Bypasses env var and settings. |

package/.pi/prompts/save.md

@@ -0,0 +1,16 @@
+---
+description: Save the current conversation or a specific insight into the wiki vault as a structured note.
+---
+
+Read the `save` skill. Then run the save workflow for this conversation.
+
+Usage:
+- `/save` — analyze the full conversation and save the most valuable content
+- `/save [name]` — save with a specific note title (skip the naming question)
+- `/save session` — save a complete session summary
+- `/save concept [name]` — explicitly save as a concept page
+- `/save decision [name]` — explicitly save as a decision record
+
+If no vault is set up yet, say: "No wiki vault found. Run /wiki first to set one up."
+
+Check if a page with the same name already exists. If it does, offer to update it instead of creating a duplicate.

package/.pi/prompts/wiki-autoresearch.md

@@ -0,0 +1,19 @@
+---
+description: Run an autonomous research loop on a topic. Searches the web, synthesizes findings, and files everything into the wiki as structured pages.
+argument-hint: "[topic]"
+---
+
+Read the `wiki-autoresearch` skill. Then run the research loop.
+
+Invocation arguments:
+- Raw arguments from command: `$ARGUMENTS`
+- If `$ARGUMENTS` is non-empty: treat it as explicit topic and use it verbatim.
+- If `$ARGUMENTS` is empty: ask exactly "What topic should I research?"
+
+Before starting, read `skills/wiki-autoresearch/references/program.md` to load the research constraints and objectives.
+
+If no vault is set up yet, say: "No wiki vault found. Run /wiki first to set one up."
+
+After research is complete, update wiki/index.md, wiki/log.md, and wiki/hot.md.
+
+Report how many pages were created and what the key findings are.

package/.pi/prompts/wiki.md

@@ -0,0 +1,23 @@
+---
+description: Bootstrap or check the claude-obsidian wiki vault. Reads the wiki skill and runs setup workflow.
+---
+
+Read the `wiki` skill. Then run the setup workflow:
+
+1. Check if Obsidian is installed. If not, offer to install it (see `skills/wiki/references/plugins.md`).
+2. Check if this directory has a vault (look for `.obsidian/` folder). If yes, report current vault state.
+3. Check if the MCP server is configured (`claude mcp list`). If not, ask if the user wants to set it up.
+4. Ask ONE question: "What is this vault for?"
+
+Then build the entire wiki structure based on the answer. Don't ask more questions. Scaffold it, show what was created, and ask: "Want to adjust anything before we start?"
+
+Examples of what the user might say:
+- "Map the architecture of github.com/org/repo"
+- "Build a sitemap and content analysis for example.com"
+- "Track my SaaS business — product, customers, metrics, roadmap"
+- "Research project on [topic] — papers, concepts, open questions"
+- "Personal second brain — health, goals, learning, projects"
+- "Organize my YouTube channel — transcripts, topics, tools mentioned"
+- "Executive assistant brain — meetings, tasks, business context"
+
+If the vault is already set up, skip to checking what has been ingested recently and offering to continue where things left off.