feat: render v2 — reset default, additive re-render, consolidated skills

Makefile

@@ -27,6 +27,7 @@ snapshot:
 demo: build
 	bash demo-setup.sh
 	vhs demo.tape
+	vhs demo-status.tape
 
 gendocs:
 	go run ./cmd/gendocs

README.md

@@ -4,9 +4,7 @@
 
 CLI for managing multi-repo development workspaces using git worktrees.
 
-Working across multiple repos means repetitive setup, scattered branches, and cleanup debt. Flow uses a YAML state file to define which repos and branches belong together, then materializes the workspace with bare clone caching and git worktrees.
-
-![flow demo](demo.gif)
+Working across multiple repos means repetitive setup, scattered branches, and cleanup debt. ${\color{cyan}\texttt{flow}}$ uses a YAML state file to define which repos and branches belong together, then materializes the workspace with bare clone caching and git worktrees.
 
 ## Features
 
@@ -16,13 +14,25 @@ Working across multiple repos means repetitive setup, scattered branches, and cl
 
 🤖 **AI agent integration** — Generate shared context files and agent instructions across repos so your AI tools have the right skills and knowledge from the start.
 
-## Why flow?
+## Why ${\color{cyan}\texttt{flow}}$?
 
 AI agents work best when they have deterministic tools instead of freeform instructions. Asking an agent to "set up a multi-repo workspace" produces inconsistent results — but giving it a CLI that manages YAML state files, bare clone caches, and git worktrees produces the same result every time.
 
-Flow is that deterministic layer. It gives agents (and humans) a small set of reliable commands for workspace lifecycle: create, define repos in YAML, render worktrees, check status. Agents call these tools through embedded skills rather than interpreting setup instructions on their own.
+${\color{cyan}\texttt{flow}}$ is that deterministic layer. It gives agents (and humans) a small set of reliable commands for workspace lifecycle: create, define repos in YAML, render worktrees, check status. Agents call these tools through embedded skills rather than interpreting setup instructions on their own.
+
+Beyond workspace creation, ${\color{cyan}\texttt{flow}}$ centralizes agent skills across all your workspaces and lets you check the status of many workstreams in parallel. It is not opinionated about which agent or editor you use — configure Claude, Cursor, or anything else in a single config file. Everything is customizable: status checks, agent skills, workspace structure.
+
+### Describe the problem, ${\color{cyan}\texttt{flow}}$ does the rest
+
+`flow init` creates a workspace pre-configured with agent instructions and skills, then launches you directly into your preferred agent — Claude, Cursor, or anything else you configure. The agent reads the embedded flow skill to understand how workspaces work, plus any custom skills you've added (like resolving repos by friendly names). Describe what you're working on and the agent handles the rest: editing the state file, rendering worktrees, and getting everything ready to go.
+
+![flow init](demo.gif)
+
+### Track all your workstreams in one place
 
-Beyond workspace creation, flow centralizes agent skills across all your workspaces and lets you check the status of many workstreams in parallel. It is not opinionated about which agent or editor you use — configure Claude, Cursor, or anything else in a single config file. Everything is customizable: status checks, agent skills, workspace structure.
+`flow status` gives you a live view across all active workspaces. Statuses are fully customizable — they're just one-line shell commands in a config file, not hardcoded logic. Have your AI generate complex status checks (query GitHub PRs, check CI, inspect git state) and hide them behind simple labels. Nothing in ${\color{cyan}\texttt{flow}}$ is opinionated; anything that could be is configurable.
+
+![flow status](demo-status.gif)
 
 ## Getting started
 
@@ -51,30 +61,19 @@ make install
 
 ### Create a workspace and start working
 
-Flow creates a workspace and launches your configured agent. Describe what you're working on and the agent handles the rest.
+${\color{cyan}\texttt{flow}}$ creates a workspace and launches your configured agent. Describe what you're working on and the agent handles the rest.
 
 ```bash
-flow init my-project
+flow init
 ```
 
 The agent reads its embedded skills to edit `state.yaml`, run `flow render`, and begin working in the repos.
 
-### Manual workflow
-
-Use `--no-exec` to skip the agent launch and set things up yourself:
-
-```bash
-flow init my-project --no-exec
-flow edit state my-project    # add repos and branches
-flow render my-project        # clone repos, create worktrees
-flow exec my-project          # launch agent manually
-```
-
 See the [spec reference](docs/specs/) for YAML file schemas and the [command reference](docs/commands/) for all commands.
 
 ## How it works
 
-Flow stores everything under `~/.flow` (override with `$FLOW_HOME`):
+${\color{cyan}\texttt{flow}}$ stores everything under `~/.flow` (override with `$FLOW_HOME`):
 
 ```
 ~/.flow/
@@ -84,8 +83,7 @@ Flow stores everything under `~/.flow` (override with `$FLOW_HOME`):
 │   └── claude/
 │       ├── CLAUDE.md                   # Shared agent instructions
 │       └── skills/
-│           ├── flow-cli/SKILL.md       # Built-in: workspace management
-│           ├── workspace-structure/SKILL.md  # Built-in: directory layout
+│           ├── flow/SKILL.md           # Built-in: workspace management
 │           └── find-repo/SKILL.md      # Your own custom skill
 ├── workspaces/
 │   └── calm-delta/                     # Workspace ID
@@ -94,7 +92,7 @@ Flow stores everything under `~/.flow` (override with `$FLOW_HOME`):
 │       ├── CLAUDE.md                   # Generated workspace context
 │       ├── .claude/
 │       │   ├── CLAUDE.md → agents/claude/CLAUDE.md
-│       │   └── skills → agents/claude/skills/
+│       │   └── skills/                # Consolidated from shared + repo skills
 │       ├── vpc-service/                # Worktree
 │       └── subnet-manager/             # Worktree
 └── repos/
@@ -105,13 +103,13 @@ Flow stores everything under `~/.flow` (override with `$FLOW_HOME`):
 
 Bare clones are shared across workspaces. Worktrees are cheap — they share the object store with the bare clone, so multiple workspaces pointing at the same repo don't duplicate data.
 
-Flow ships two built-in skills (`flow-cli` and `workspace-structure`) and preserves any custom skills you add to the same directory. Run `flow reset skills` to update the built-in skills without touching your own.
+${\color{cyan}\texttt{flow}}$ ships a built-in `flow` skill and consolidates skills from all repos into each workspace's `.claude/skills/` directory on render. Add your own skills to the shared directory or to individual repos. Run `flow reset skills` to update the built-in skill without touching your own.
 
 See the [spec reference](docs/specs/) for YAML file schemas and the [command reference](docs/commands/) for usage, flags, and GIF demos.
 
 ## Customization
 
-Flow stores everything under `~/.flow` (override with `$FLOW_HOME`). Edit these files to customize your setup:
+${\color{cyan}\texttt{flow}}$ stores everything under `~/.flow` (override with `$FLOW_HOME`). Edit these files to customize your setup:
 
 | Command | What it configures |
 |---------|-------------------|

demo-setup.sh

@@ -8,53 +8,41 @@ FLOW="$(pwd)/flow"
 rm -rf /tmp/flow-demo /tmp/demo
 mkdir -p /tmp/demo "$FLOW_HOME/workspaces" "$FLOW_HOME/repos"
 
-# --- Create a fake claude binary that simulates an interactive agent session ---
+# --- Configure flow agent ---
+# A launcher script writes .claude/settings.json to pre-approve tools before
+# launching claude. This avoids permission dialogs during the demo.
 
-mkdir -p /tmp/flow-demo/bin
-cat > /tmp/flow-demo/bin/claude <<'SCRIPT'
+cat > /tmp/flow-demo/launch-claude.sh <<'LAUNCHER'
 #!/bin/bash
-DIM='\033[2m'
-CYAN='\033[36m'
-GREEN='\033[32m'
-BOLD='\033[1m'
-RESET='\033[0m'
-
-echo ""
-echo -e "  ${DIM}[mocked agent session]${RESET}"
-echo ""
-echo -e "  ${DIM}Flow includes skills that teach your agent to manage workspaces.${RESET}"
-echo -e "  ${DIM}Add your own skills for repo discovery, PR lookup, or any custom workflow.${RESET}"
-echo -e "  ${DIM}Paste a Slack thread, dictate a bug report — the agent handles the rest.${RESET}"
-echo ""
-echo -e "  ${BOLD}Enter your prompt:${RESET}"
-echo ""
-printf "  > "
-read -r task
-
-echo ""
-sleep 0.5
-echo -e "  ${CYAN}● Reading skill: flow-cli${RESET}"
-sleep 0.6
-echo -e "  ${CYAN}● Reading skill: find-repo${RESET}"
-echo -e "      ${DIM}→ github.com/acme/web${RESET}"
-sleep 0.6
-echo -e "  ${CYAN}● Editing state.yaml${RESET}"
-echo -e "      ${DIM}name: fix-dashboard-charts${RESET}"
-echo -e "      ${DIM}repo: github.com/acme/web @ fix/dashboard-charts${RESET}"
-sleep 0.6
-echo -e "  ${CYAN}● Running flow render...${RESET}"
-sleep 0.8
-echo ""
-echo -e "  ${GREEN}✓ Workspace ready${RESET}"
-echo ""
-echo -e "  ${CYAN}● Analyzing web/src/components/Charts.tsx...${RESET}"
-sleep 0.8
-echo ""
-printf "  > "
-read -r cmd
-echo ""
-SCRIPT
-chmod +x /tmp/flow-demo/bin/claude
+echo '{"permissions":{"allow":["Bash","Edit","Write","Read","Glob","Grep","Skill"]}}' > .claude/settings.json
+exec claude --model haiku
+LAUNCHER
+chmod +x /tmp/flow-demo/launch-claude.sh
+
+cat > "$FLOW_HOME/config.yaml" <<YAML
+apiVersion: flow/v1
+kind: Config
+spec:
+    agents:
+        - name: claude
+          exec: /tmp/flow-demo/launch-claude.sh
+          default: true
+YAML
+
+# --- Pre-trust the demo directory so Claude skips the workspace trust dialog ---
+
+CLAUDE_JSON="$HOME/.claude.json"
+if [ -f "$CLAUDE_JSON" ]; then
+  python3 -c "
+import json, sys
+with open('$CLAUDE_JSON') as f:
+    config = json.load(f)
+config.setdefault('projects', {})
+config['projects']['/tmp/flow-demo'] = {'hasTrustDialogAccepted': True}
+with open('$CLAUDE_JSON', 'w') as f:
+    json.dump(config, f, indent=2)
+"
+fi
 
 # --- Create local git repos with feature branches ---
 
@@ -67,29 +55,36 @@ create_repo() {
   echo "package main" > "$dir/main.go"
   git -C "$dir" add .
   git -C "$dir" commit -q -m "initial commit"
-  git -C "$dir" checkout -q -b "$branch"
+  git -C "$dir" checkout -q -b "$branch" 2>/dev/null || true
   echo "// $msg" >> "$dir/$file"
   git -C "$dir" add .
   git -C "$dir" commit -q -m "$msg"
 }
 
-create_repo "app"     "feature/ipv6"     "main.go"   "add ipv6 support"
-create_repo "api"     "feat/auth"        "main.go"   "add auth endpoints"
-create_repo "docs"    "update/guides"    "README.md" "update setup guide"
-create_repo "web"     "feat/dashboard"   "main.go"   "add dashboard page"
-create_repo "billing" "feat/billing-v2"  "main.go"   "billing v2 migration"
-create_repo "gateway" "feat/rate-limits" "main.go"   "add rate limiting"
-create_repo "config"  "feat/env-vars"    "main.go"   "add env var support"
+create_repo "app"            "feature/ipv6"     "main.go"   "add ipv6 support"
+create_repo "api"            "feat/auth"        "main.go"   "add auth endpoints"
+create_repo "docs"           "update/guides"    "README.md" "update setup guide"
+create_repo "web"            "feat/dashboard"   "main.go"   "add dashboard page"
+create_repo "billing"        "feat/billing-v2"  "main.go"   "billing v2 migration"
+create_repo "gateway"        "feat/rate-limits" "main.go"   "add rate limiting"
+create_repo "config"         "feat/env-vars"    "main.go"   "add env var support"
+create_repo "apps"           "main"             "main.go"   "initial app setup"
+create_repo "infrastructure" "main"             "main.go"   "initial infra setup"
 
 # --- Pre-populate bare clone cache for realistic URLs ---
 
-for name in app api docs web billing gateway config; do
+for name in app api docs web billing gateway config apps infrastructure; do
   bare="$FLOW_HOME/repos/github.com/acme/${name}.git"
   mkdir -p "$(dirname "$bare")"
   git clone --bare "/tmp/demo/$name" "$bare" -q
   # Add fetch refspec so worktrees can resolve origin/* refs
   git -C "$bare" config remote.origin.fetch "+refs/heads/*:refs/remotes/origin/*"
   git -C "$bare" fetch -q origin
+
+  # Also create under SSH URL path (in case Claude uses git@ format)
+  ssh_bare="$FLOW_HOME/repos/git@github.com:acme/${name}.git"
+  mkdir -p "$(dirname "$ssh_bare")"
+  cp -a "$bare" "$ssh_bare"
 done
 
 # --- Create and render pre-existing workspaces ---
@@ -184,7 +179,7 @@ YAML
 )"
 
 # Render all workspaces
-# Use --reset=false to skip interactive prompt when branches already exist
+# Use --reset=false to use existing branches as-is instead of creating fresh from base
 $FLOW render bold-creek --reset=false
 $FLOW render swift-pine --reset=false
 $FLOW render calm-ridge --reset=false

demo-status.tape

@@ -0,0 +1,52 @@
+# Regenerate: make demo
+Output demo-status.gif
+
+Set Shell "bash"
+Set FontFamily "Menlo"
+Set FontSize 16
+Set Width 1100
+Set Height 650
+Set Padding 20
+Set Theme "Catppuccin Mocha"
+Set TypingSpeed 30ms
+Set PlaybackSpeed 1
+Set LetterSpacing 0
+Set LineHeight 1.2
+
+# --- Hidden env setup ---
+Hide
+
+Type `export FLOW_HOME=/tmp/flow-demo/.flow`
+Enter
+Sleep 200ms
+
+Type `export PATH="$(pwd):$PATH"`
+Enter
+Sleep 200ms
+
+# Swap in local-only status spec (no gh needed for demo)
+Type `cp "$FLOW_HOME/status-local.yaml" "$FLOW_HOME/status.yaml"`
+Enter
+Sleep 200ms
+
+Type@0ms `clear`
+Enter
+Sleep 500ms
+
+Show
+
+# === See all workstreams at a glance ===
+
+Sleep 500ms
+
+Type "flow status"
+Sleep 500ms
+Enter
+Sleep 5s
+
+Type "flow status api-refactor"
+Sleep 500ms
+Enter
+Sleep 4s
+
+Sleep 4s

demo.tape

@@ -4,8 +4,8 @@ Output demo.gif
 Set Shell "bash"
 Set FontFamily "Menlo"
 Set FontSize 16
-Set Width 900
-Set Height 550
+Set Width 1100
+Set Height 650
 Set Padding 20
 Set Theme "Catppuccin Mocha"
 Set TypingSpeed 30ms
@@ -20,63 +20,37 @@ Type `export FLOW_HOME=/tmp/flow-demo/.flow`
 Enter
 Sleep 200ms
 
-Type `export PATH="/tmp/flow-demo/bin:$(pwd):$PATH"`
+Type `export PATH="$(pwd):$PATH"`
 Enter
 Sleep 200ms
 
-# Swap in local-only status spec (no gh needed for demo)
-Type `cp "$FLOW_HOME/status-local.yaml" "$FLOW_HOME/status.yaml"`
-Enter
-Sleep 200ms
-
-# Clear screen and print first section comment (hidden from recording)
-Type@0ms `clear && printf '\033[2;3m# create a workspace — agent launches automatically\033[0m\n'`
+Type@0ms `clear`
 Enter
 Sleep 500ms
 
 Show
 
-# === Create a workspace — agent launches automatically ===
-
-Sleep 500ms
-
-Type "flow init dashboard"
-Sleep 500ms
-Enter
-Sleep 3s
+# === flow init — launches Claude automatically ===
 
-# The fake claude is now running and showing its prompt.
-# Type the task description into the agent.
-Type "dashboard charts broke after the auth migration, check the web frontend"
 Sleep 500ms
-Enter
-Sleep 5s
 
-# Agent has finished setting up. Exit the session.
-Type "/exit"
+Type "flow init"
 Sleep 500ms
 Enter
-Sleep 2s
 
-# === See all workstreams at a glance ===
+# Wait for Claude to launch and show the workspace trust dialog
+Sleep 3s
 
-# Clear screen and print second section comment (hidden from recording)
+# Accept the workspace trust dialog
 Hide
-Type@0ms `clear && printf '\033[2;3m# see all workstreams at a glance\033[0m\n'`
 Enter
-Sleep 500ms
 Show
+Sleep 3s
 
-Sleep 500ms
-
-Type "flow status"
-Sleep 500ms
-Enter
-Sleep 5s
-
-Type "flow status api-refactor"
+# Type the task into Claude
+Type "Our auth migration broke prod — API errors in github.com/acme/apps, wrong docs in github.com/acme/docs, and infra config drift in github.com/acme/infrastructure. Set up a workspace for this, then ask me what to look at first."
 Sleep 500ms
 Enter
-Sleep 4s
 
-Sleep 4s
+# Wait for Claude to read skills, edit state.yaml, run flow render
+Sleep 40s