update README, skill definitions, and website layout

Rewrite README with comprehensive AI-DLC documentation covering
installation, terminology, unit model, workflows, and conventions.
Update skill SKILL.md files for construct, elaborate, fail, operate,
reflect, resume, and setup with refined instructions. Fix website
layout.tsx and inject-context hook.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: jwaldrip

README.md

@@ -554,7 +554,7 @@ fi
 # Display Agent Teams status if enabled
 if [ -n "${CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS:-}" ]; then
   TEAM_NAME="ai-dlc-${INTENT_SLUG}"
-  TEAM_CONFIG="$HOME/.claude/teams/${TEAM_NAME}/config.json"
+  TEAM_CONFIG="${CLAUDE_CONFIG_DIR}/teams/${TEAM_NAME}/config.json"
   if [ -f "$TEAM_CONFIG" ]; then
     echo "### Agent Teams"
     echo ""

plugin/hooks/inject-context.sh

@@ -154,7 +154,10 @@ if ! grep -q '\.ai-dlc/worktrees/' "${REPO_ROOT}/.gitignore" 2>/dev/null; then
 fi
 
 if [ ! -d "$INTENT_WORKTREE" ]; then
-  git worktree add -B "$INTENT_BRANCH" "$INTENT_WORKTREE"
+  # Always branch off the default branch
+  source "${CLAUDE_PLUGIN_ROOT}/lib/config.sh"
+  DEFAULT_BRANCH=$(resolve_default_branch "auto" "$REPO_ROOT")
+  git worktree add -B "$INTENT_BRANCH" "$INTENT_WORKTREE" "$DEFAULT_BRANCH"
 fi
 
 cd "$INTENT_WORKTREE"
@@ -372,7 +375,7 @@ Check if the team already exists before creating:
 
 ```bash
 TEAM_NAME="ai-dlc-${INTENT_SLUG}"
-TEAM_CONFIG="$HOME/.claude/teams/${TEAM_NAME}/config.json"
+TEAM_CONFIG="${CLAUDE_CONFIG_DIR}/teams/${TEAM_NAME}/config.json"
 ```
 
 If team config does NOT exist:
@@ -421,7 +424,10 @@ UNIT_BRANCH="ai-dlc/${INTENT_SLUG}/${UNIT_SLUG}"
 WORKTREE_PATH="${REPO_ROOT}/.ai-dlc/worktrees/${INTENT_SLUG}-${UNIT_SLUG}"
 
 if [ ! -d "$WORKTREE_PATH" ]; then
-  git worktree add -B "$UNIT_BRANCH" "$WORKTREE_PATH"
+  # Always branch off the default branch
+  source "${CLAUDE_PLUGIN_ROOT}/lib/config.sh"
+  DEFAULT_BRANCH=$(resolve_default_branch "auto" "$REPO_ROOT")
+  git worktree add -B "$UNIT_BRANCH" "$WORKTREE_PATH" "$DEFAULT_BRANCH"
 fi
 ```
 
@@ -1052,7 +1058,10 @@ WORKTREE_PATH="${REPO_ROOT}/.ai-dlc/worktrees/${intentSlug}-${UNIT_SLUG}"
 
 # Create worktree if it doesn't exist
 if [ ! -d "$WORKTREE_PATH" ]; then
-  git worktree add -B "$UNIT_BRANCH" "$WORKTREE_PATH"
+  # Always branch off the default branch
+  source "${CLAUDE_PLUGIN_ROOT}/lib/config.sh"
+  DEFAULT_BRANCH=$(resolve_default_branch "auto" "$REPO_ROOT")
+  git worktree add -B "$UNIT_BRANCH" "$WORKTREE_PATH" "$DEFAULT_BRANCH"
 fi
 ```
 

plugin/skills/construct/SKILL.md

@@ -166,6 +166,10 @@ Agent({
 
 Spawn one subagent per design file, in parallel with codebase Explore agents. When results return:
 - Append to `discovery.md` under `## Design Analysis: {file name}`
+- **Commit immediately:**
+  ```bash
+  git add .ai-dlc/${INTENT_SLUG}/discovery.md && git commit -m "elaborate(${INTENT_SLUG}): discover design analysis for {file name}"
+  ```
 - If no design MCP tools are discoverable, the subagent reports unavailability — log a warning and continue without design analysis
 
 5. **UI Mockups**: If the intent involves user-facing interfaces (frontend, CLI, TUI, etc.), generate mockups for every distinct screen or view. This step is **mandatory** for any intent with a UI component:
@@ -231,6 +235,11 @@ Spawn one subagent per design file, in parallel with codebase Explore agents. Wh
    - {UI region} ← {domain entity}.{field}
    ```
 
+   **Commit immediately after each mockup:**
+   ```bash
+   git add .ai-dlc/${INTENT_SLUG}/discovery.md && git commit -m "elaborate(${INTENT_SLUG}): discover UI mockup for {View Name}"
+   ```
+
    **Skip mockups only if:** the intent has no user-facing interface (pure backend, API, data pipeline, infrastructure, etc.).
 
 **Spawn multiple research paths in parallel.** Don't serialize explorations that are independent — launch all of them at once and synthesize when results return.
@@ -253,6 +262,11 @@ After each significant finding (API schema mapped, codebase pattern identified,
 - `## Technology Choice: {name}` — For greenfield/early projects: technology selection rationale
 - `## Reference Implementation: {name}` — For greenfield/early projects: external reference implementations or prior art informing the design
 
+**Commit immediately after each append to discovery.md:**
+```bash
+git add .ai-dlc/${INTENT_SLUG}/discovery.md && git commit -m "elaborate(${INTENT_SLUG}): discover {topic}"
+```
+
 **After appending to discovery.md, keep only a brief summary in your context** — the full details are safely on disk and will be available to builders. This is the key benefit: your context stays lean for continued exploration while nothing is lost.
 
 **CRITICAL**: Do not summarize or skip exploration. The exploration results directly determine whether the spec is accurate. If you explore a GraphQL API, report every type. If you read source code, report the actual architecture, not your guess about it.
@@ -283,6 +297,11 @@ Structure the domain model as:
 - {description of any gap between what's needed and what's available}
 - {proposed solution for each gap}
 
+**Commit the domain model immediately after appending it to discovery.md:**
+```bash
+git add .ai-dlc/${INTENT_SLUG}/discovery.md && git commit -m "elaborate(${INTENT_SLUG}): discover domain model"
+```
+
 ---
 
 ## Step 4: Write Results
@@ -332,12 +351,21 @@ error_message: ""
 - {path to mockup 2} — {description}
 ```
 
+**Commit the results file immediately after writing:**
+```bash
+git add .ai-dlc/${INTENT_SLUG}/.briefs/elaborate-discover-results.md && git commit -m "elaborate(${INTENT_SLUG}): write discovery results"
+```
+
 ---
 
 ## Error Handling
 
 If any critical error occurs during exploration (e.g., worktree path doesn't exist, discovery.md not found):
 
 1. Write the results file with `status: error` and `error_message` describing what went wrong
-2. Include any partial findings that were gathered before the error
-3. Exit — the main elaborate skill will read the error status and handle it
+2. **Commit the error results file immediately:**
+   ```bash
+   git add .ai-dlc/${INTENT_SLUG}/.briefs/elaborate-discover-results.md && git commit -m "elaborate(${INTENT_SLUG}): write discovery results (error)"
+   ```
+3. Include any partial findings that were gathered before the error
+4. Exit — the main elaborate skill will read the error status and handle it

plugin/skills/elaborate-discover/SKILL.md

@@ -252,7 +252,10 @@ if ! grep -q '\.ai-dlc/worktrees/' "${REPO_ROOT}/.gitignore" 2>/dev/null; then
   exit 1
 fi
 
-git worktree add -B "$INTENT_BRANCH" "$INTENT_WORKTREE"
+# Always branch off the default branch
+source "${CLAUDE_PLUGIN_ROOT}/lib/config.sh"
+DEFAULT_BRANCH=$(resolve_default_branch "auto" "$REPO_ROOT")
+git worktree add -B "$INTENT_BRANCH" "$INTENT_WORKTREE" "$DEFAULT_BRANCH"
 cd "$INTENT_WORKTREE"
 ```
 
@@ -340,6 +343,13 @@ provider_config: {JSON of PROVIDERS object}
 {absolute path to discovery.md}
 ```
 
+Commit the discovery brief immediately after writing it:
+
+```bash
+git add .ai-dlc/${INTENT_SLUG}/.briefs/elaborate-discover.md
+git commit -m "elaborate(${INTENT_SLUG}): write discovery brief"
+```
+
 ### Step 3: Invoke discovery subagent
 
 ```
@@ -1165,6 +1175,13 @@ design_provider_type: {DESIGN_TYPE or empty}
 {Abbreviated domain model from intent.md}
 ```
 
+Commit the wireframes brief immediately after writing it:
+
+```bash
+git add .ai-dlc/${INTENT_SLUG}/.briefs/elaborate-wireframes.md
+git commit -m "elaborate(${INTENT_SLUG}): write wireframes brief"
+```
+
 ### Step 4: Invoke wireframes subagent
 
 ```
@@ -1181,6 +1198,13 @@ Read `.ai-dlc/${INTENT_SLUG}/.briefs/elaborate-wireframes-results.md`.
 - If `status: error` — report the error to the user and discuss how to proceed
 - If `status: success` — list the generated wireframes for the user
 
+Commit the wireframe artifacts and results brief:
+
+```bash
+git add .ai-dlc/${INTENT_SLUG}/mockups/ .ai-dlc/${INTENT_SLUG}/.briefs/elaborate-wireframes-results.md
+git commit -m "elaborate(${INTENT_SLUG}): generate frontend wireframes"
+```
+
 ### Step 6: Product review gate
 
 Present all generated wireframes to product for review using `AskUserQuestion`:
@@ -1201,7 +1225,11 @@ Present all generated wireframes to product for review using `AskUserQuestion`:
 ```
 
 - **Approved**: Proceed to Phase 6.5.
-- **Needs revision**: Discuss feedback, update the wireframe HTML files directly, and re-present for review. Loop until approved.
+- **Needs revision**: Discuss feedback, update the wireframe HTML files directly, commit the revision, and re-present for review. Loop until approved.
+  ```bash
+  git add .ai-dlc/${INTENT_SLUG}/mockups/
+  git commit -m "elaborate(${INTENT_SLUG}): revise wireframes"
+  ```
 - **Skip wireframes**: Delete the `mockups/` directory, remove `wireframe:` fields from unit frontmatter, and proceed to Phase 6.5.
 
 ---
@@ -1260,6 +1288,13 @@ plugin_root: {CLAUDE_PLUGIN_ROOT}
 - **Wireframe:** {wireframe path from frontmatter, if any}
 ```
 
+Commit the ticket sync brief immediately after writing it:
+
+```bash
+git add .ai-dlc/${INTENT_SLUG}/.briefs/elaborate-ticket-sync.md
+git commit -m "elaborate(${INTENT_SLUG}): write ticket sync brief"
+```
+
 ### Step 3: Invoke ticket sync subagent
 
 ```
@@ -1276,6 +1311,13 @@ Read `.ai-dlc/${INTENT_SLUG}/.briefs/elaborate-ticket-sync-results.md`.
 - If `status: error` — report the errors to the user. If `validation_passed: false`, the subagent already retried. Log the failures and proceed to Phase 7 (never block elaboration on ticket sync failure)
 - If `status: success` — log the epic key and ticket keys, confirm validation passed, proceed to Phase 7
 
+Commit the ticket sync artifacts (updated intent.md and unit files with ticket keys, plus the results brief):
+
+```bash
+git add .ai-dlc/${INTENT_SLUG}/
+git commit -m "elaborate(${INTENT_SLUG}): sync tickets to provider"
+```
+
 ---
 
 ## Phase 7: Handoff

plugin/skills/elaborate/SKILL.md

@@ -61,6 +61,17 @@ REASON="Reviewer found issues: [describe issues]"
 han keep save blockers.md "$REASON"
 ```
 
+### Step 3a: Commit Blocker Documentation
+
+If any blocker documentation was written to the working tree (not han keep), commit immediately:
+
+```bash
+if [ -n "$(git status --porcelain)" ]; then
+  git add -A
+  git commit -m "ai-dlc(${INTENT_SLUG}): document blocker"
+fi
+```
+
 ### Step 4: Update State
 
 ```bash

plugin/skills/fail/SKILL.md

@@ -131,6 +131,17 @@ if [ -n "$TASK_COMMAND" ]; then
 fi
 ```
 
+### Step 4a: Commit Agent Task Changes
+
+If any agent-owned task execution modified files in the working tree, commit immediately after the task completes:
+
+```bash
+if [ -n "$(git status --porcelain)" ]; then
+  git add -A
+  git commit -m "operate(${INTENT_SLUG}): execute ${TASK_NAME}"
+fi
+```
+
 ### Step 5: Handle Human-Owned Tasks
 
 For tasks where `owner: human`:

plugin/skills/operate/SKILL.md

@@ -99,7 +99,7 @@ Session transcripts are stored as JSONL files. Find sessions associated with thi
 
 ```bash
 # Claude Code session logs location
-PROJECT_HASH_DIR="$HOME/.claude/projects/"
+PROJECT_HASH_DIR="${CLAUDE_CONFIG_DIR}/projects/"
 
 # Find the project hash directory for the current repo
 REPO_ROOT=$(git rev-parse --show-toplevel 2>/dev/null || pwd)
@@ -228,6 +228,12 @@ status: complete
 {What v2 should focus on, if the user chooses to iterate}
 ```
 
+Commit the reflection artifact immediately after writing:
+
+```bash
+git add .ai-dlc/${INTENT_SLUG}/reflection.md && git commit -m "reflect(${INTENT_SLUG}): capture reflection"
+```
+
 ### Step 4b: Produce settings-recommendations.md
 
 Based on session analysis and execution patterns, produce concrete settings changes. Write to `.ai-dlc/{intent-slug}/settings-recommendations.md`:
@@ -275,6 +281,12 @@ Example:
 - Success criteria should specify test coverage thresholds numerically (units with "adequate coverage" criteria caused 2x reviewer cycles vs units with "80% line coverage")
 ```
 
+Commit the settings recommendations artifact immediately after writing:
+
+```bash
+git add .ai-dlc/${INTENT_SLUG}/settings-recommendations.md && git commit -m "reflect(${INTENT_SLUG}): document settings recommendations"
+```
+
 ### Step 5: Present Findings for Validation
 
 Output the reflection summary and settings recommendations. Ask the user to:
@@ -394,6 +406,12 @@ If user chooses to close:
 - {Process improvement that applies broadly}
 ```
 
+Commit the learnings immediately after writing:
+
+```bash
+git add .claude/memory/learnings.md && git commit -m "reflect(${INTENT_SLUG}): capture learnings"
+```
+
 3. **Sync to HAIKU organizational memory** (if workspace configured):
 
 ```bash

plugin/skills/reflect/SKILL.md

@@ -127,7 +127,10 @@ if ! grep -q '\.ai-dlc/worktrees/' "${REPO_ROOT}/.gitignore" 2>/dev/null; then
 fi
 
 if [ ! -d "$INTENT_WORKTREE" ]; then
-  git worktree add -B "$INTENT_BRANCH" "$INTENT_WORKTREE"
+  # Always branch off the default branch
+  source "${CLAUDE_PLUGIN_ROOT}/lib/config.sh"
+  DEFAULT_BRANCH=$(resolve_default_branch "auto" "$REPO_ROOT")
+  git worktree add -B "$INTENT_BRANCH" "$INTENT_WORKTREE" "$DEFAULT_BRANCH"
 fi
 cd "$INTENT_WORKTREE"
 ```
@@ -151,7 +154,7 @@ When `CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS` is enabled:
 ```bash
 AGENT_TEAMS_ENABLED="${CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS:-}"
 TEAM_NAME="ai-dlc-${SLUG}"
-TEAM_CONFIG="$HOME/.claude/teams/${TEAM_NAME}/config.json"
+TEAM_CONFIG="${CLAUDE_CONFIG_DIR}/teams/${TEAM_NAME}/config.json"
 ```
 
 If `AGENT_TEAMS_ENABLED` is set:

plugin/skills/resume/SKILL.md

@@ -229,15 +229,25 @@ For each **confirmed provider**, offer to customize how AI-DLC interacts with it
      {contents of built-in default, body only (after frontmatter)}
      ```
    - Tell the user: "Created `.ai-dlc/providers/{type}.md` with defaults as a starting point. Edit this file to customize how AI-DLC interacts with {type} for this project."
+   - Commit the new override file immediately:
+     ```bash
+     git add .ai-dlc/providers/{type}.md && git commit -m "ai-dlc: configure {type} provider"
+     ```
 
 5. **If "Use defaults"** → skip, no file created. The built-in defaults apply automatically.
 
 6. **If an override file already exists** from a previous `/setup` run, change the question to:
    - "Project override exists at `.ai-dlc/providers/{type}.md`. Want to keep it, reset to defaults, or remove it?"
    - Options:
      - **"Keep as-is"** — Don't touch the existing override
-     - **"Reset to defaults"** — Overwrite with current built-in defaults
-     - **"Remove override"** — Delete the file, revert to built-in defaults only
+     - **"Reset to defaults"** — Overwrite with current built-in defaults, then commit:
+       ```bash
+       git add .ai-dlc/providers/{type}.md && git commit -m "ai-dlc: reset {type} provider to defaults"
+       ```
+     - **"Remove override"** — Delete the file, revert to built-in defaults only, then commit the removal:
+       ```bash
+       git rm .ai-dlc/providers/{type}.md && git commit -m "ai-dlc: remove {type} provider override"
+       ```
 
 **Skip** this phase for any provider category that has no confirmed provider.
 
@@ -307,6 +317,11 @@ Rules:
 
 3. Write the file using the `Write` tool to `.ai-dlc/settings.yml`.
 
+4. Commit the settings file immediately:
+   ```bash
+   git add .ai-dlc/settings.yml && git commit -m "ai-dlc: configure project settings"
+   ```
+
 ---
 
 ## Phase 7: Confirmation