bigpowers 1.0.0 → 1.1.0

package/release-branch/SKILL.md

@@ -1,5 +1,6 @@
 ---
 name: release-branch
+model: haiku
 description: Make the merge/PR/keep/discard decision for a feature branch, verify coverage gates, create the PR with gh, and clean up the worktree. Use when a feature is done and ready to ship, or when user says "release", "merge", or "open a PR".
 ---
 
@@ -7,7 +8,18 @@ description: Make the merge/PR/keep/discard decision for a feature branch, verif
 
 > **HARD GATE** — Do NOT merge or release if tests fail or if coverage gates are not met. If the branch is red, return to `develop-tdd` to fix regressions or add missing tests before proceeding.
 
-Finalize a completed feature branch: verify coverage gates, open a PR, and clean up the worktree.
+Finalize a completed feature branch: verify coverage gates, integrate onto `main`, and clean up the worktree.
+
+## Integrate mode
+
+Choose mode from project profile or explicit user request:
+
+| Mode | When | Ship path |
+|------|------|-----------|
+| **solo-local** | `profiles/solo-git.md` or `specs/WORKFLOW-solo-git.md` active | `bash scripts/land-branch.sh <branch> "<conventional message>"` — no PR |
+| **team-pr** | Default, or collaboration / remote CI required | `gh pr create` → `gh pr merge --squash` (§6–7) |
+
+If unsure and working alone, prefer **solo-local**.
 
 ## Process
 
@@ -49,22 +61,37 @@ Confirm:
 - [ ] No secrets, credentials, or personal data in the diff
 - [ ] CONVENTIONS.md compliance across all changes
 
-### 5. Decision
+### 4. Decision
 
 Present the user with the options:
 
 | Option | When to choose |
 |--------|---------------|
-| **Open PR for Release** | Feature is complete, tests pass, ready to trigger automated release |
+| **Release (solo-local)** | Feature complete; solo profile active — land via script |
+| **Open PR for Release** | Feature complete; team-pr mode or remote CI gate needed |
 | **Keep branch** | More work needed; preserve for later |
 | **Discard** | Approach was wrong; start over |
 
-### 6. Create PR (Triggers Automated Release)
+### 5. Solo-local integrate
+
+From the **primary** repository root (not the feature worktree):
+
+1. Run `commit-message` to produce the squash commit subject.
+2. Land:
+
+```bash
+bash scripts/land-branch.sh <task-slug> "feat(scope): description"
+```
+
+The script will: update `main`, squash-merge, commit, push, remove worktree, delete branch, and leave you on `main`.
+
+Skip duplicate cleanup in §8 if the script succeeded.
+
+### 6. Create PR (team-pr only)
 
 The PR title is the **single source of truth** for the version bump. It MUST follow Conventional Commits.
 
 ```bash
-# Verify PR Title first
 PR_TITLE="<type>(<scope>): <description>"
 echo "$PR_TITLE" | grep -vE "^(feat|fix|docs|style|refactor|perf|test|build|ci|chore|revert)(\(.+\))?!?: .+$" && echo "❌ ERROR: PR Title must follow Conventional Commits"
 
@@ -87,25 +114,25 @@ EOF
 )"
 ```
 
-### 7. Merge (Automated)
+### 7. Merge (team-pr only)
 
-Wait for CI to pass. Merge using **Squash and Merge** to ensure the PR title becomes the commit message on `main`.
+Wait for CI to pass. Merge using **Squash and Merge** so the PR title becomes the commit message on `main`.
 
 ```bash
 gh pr merge --squash --delete-branch
 ```
 
-`semantic-release` will now automatically:
+`semantic-release` will automatically:
 1. Detect the commit on `main`.
 2. Determine the SemVer bump from the commit type.
 3. Tag the repo (e.g., `v2.1.0`).
 4. Generate release notes.
 
-### 8. Clean up worktree (if using git worktree)
+### 8. Clean up worktree (if not done by land-branch.sh)
 
 ```bash
 # From the main repo root
-git worktree prune             # clear stale metadata
+git worktree prune
 git worktree remove ../<branch-name> 2>/dev/null || true
 git branch -d <branch-name>
 ```
@@ -113,4 +140,17 @@ git branch -d <branch-name>
 - If `git worktree remove` fails due to uncommitted changes, ask the user: "There are uncommitted changes in the worktree. Force remove? (y/n)". If yes: `git worktree remove -f ../<branch-name>`.
 - If the directory `../<branch-name>` is already missing, `git worktree remove` might fail; the `|| true` ensures the process continues to branch deletion.
 
-Report: "Branch released. PR: <URL>. Worktree cleaned up."
+### 9. Return to main (primary worktree)
+
+```bash
+cd <primary-repo-root>
+git checkout main   # or master
+git status
+pwd
+```
+
+Confirm:
+- [ ] Current branch is `main` (or project default)
+- [ ] cwd is the primary repository root, not `../<task-slug>`
+
+Report: "Branch released. Integrate mode: <solo-local|team-pr>. cwd: $(pwd) on $(git branch --show-current)."

package/request-review/SKILL.md

@@ -1,5 +1,6 @@
 ---
 name: request-review
+model: opus
 description: Dispatch a fresh reviewer agent with a clean context to critique the code after audit-code passes. The reviewer has no shared state with the coding agent and gives a genuine second opinion. Use after audit-code passes, before committing, or when user wants an independent code review.
 ---
 
@@ -21,7 +22,7 @@ Write a self-contained brief for the reviewer agent. Include:
 
 - What was built (feature description, not implementation)
 - Which files changed (the diff context)
-- What `specs/` artifacts are relevant (PLAN.md, DIAGNOSIS.md, SCOPE.md)
+- What `specs/` artifacts are relevant (PLAN.md, bugs/BUG-*.md, SCOPE.md)
 - What CONVENTIONS.md requires
 - What the verify command is
 - What you're most uncertain about (where you want fresh eyes)

package/research-first/REFERENCE.md

@@ -0,0 +1,29 @@
+# Research First — Reference
+
+## Search commands
+
+```bash
+# Repo prior art
+rg -l "<keyword>" --glob '!node_modules' .
+find . -maxdepth 3 -name "SKILL.md" | xargs grep -l "<intent>"
+
+# Installed packages (if package.json exists)
+cat package.json | jq '.dependencies,.devDependencies' 2>/dev/null
+```
+
+## Registry checklist
+
+- [ ] npm / PyPI / crates.io (if applicable)
+- [ ] Existing bigpowers skill (`bash scripts/build-skill-index.sh && rg "<intent>" specs/SKILL-SEARCH-INDEX.md`)
+- [ ] Project `docs/` and `specs/adr/`
+- [ ] Official library documentation (quote one API detail)
+
+## Prior Art template
+
+```markdown
+## Prior Art
+
+| Candidate | Source | Verdict | Notes |
+|-----------|--------|---------|-------|
+| ... | ... | adopt/extend/compose/build | ... |
+```

package/research-first/SKILL.md

@@ -0,0 +1,31 @@
+---
+name: research-first
+description: Look-before-build — search registries, repo, existing skills, and web for prior art before implementing. Appends Prior Art to the spec. Use after survey-context and before elaborate-spec, when adding dependencies, or when the task may already be solved.
+model: sonnet
+---
+
+# Research First
+
+> **HARD GATE** — Do NOT implement until prior art is searched. Minimum outcome: adopt, extend, compose, or build — with evidence.
+
+## Process
+
+1. Read `specs/SCOPE.md`, `specs/RELEASE-PLAN.md`, and the current task statement.
+2. Search in order: this repo → bigpowers skills (`search-skills`) → package registries → web docs.
+3. For each candidate: note name, URL/path, fit (adopt | extend | compose | build).
+4. Append `## Prior Art` to the active spec (SCOPE.md or story in RELEASE-PLAN.md).
+
+## Outcome matrix
+
+| Verdict | Action |
+|---------|--------|
+| **adopt** | Use as-is; link in plan; no new code |
+| **extend** | Wrap or configure existing solution |
+| **compose** | Chain existing skills/modules |
+| **build** | New implementation — justify why others failed |
+
+## Verify
+
+→ verify: `grep -c "Prior Art" specs/SCOPE.md specs/RELEASE-PLAN.md 2>/dev/null | awk '{s+=$1} END {if(s>0) print "OK"; else print "MISSING"}'`
+
+See [REFERENCE.md](REFERENCE.md) for search commands and registry checklist.

package/reset-baseline/SKILL.md

@@ -0,0 +1,21 @@
+---
+name: reset-baseline
+description: Restore the project to a known clean state between agent runs or experiments. Use between benchmark runs, after a failed spike, or when user wants a clean working tree.
+model: haiku
+---
+
+# Reset Baseline
+
+> **HARD GATE** — Confirm with user before any destructive git operation. Never `reset --hard` without explicit approval.
+
+## Process
+
+1. `git status` — list uncommitted and untracked files.
+2. Ask: stash, discard, or keep each category.
+3. Safe defaults: `git stash push -u -m "reset-baseline"` for WIP; never force-push.
+4. Re-run `setup-environment` after reset.
+5. Run test baseline from `kickoff-branch` verify command.
+
+## Verify
+
+→ verify: `git status --short | wc -l | awk '{if($1==0) print "OK"; else print "DIRTY:" $1}'

package/respond-review/SKILL.md

@@ -1,5 +1,6 @@
 ---
 name: respond-review
+model: sonnet
 description: Act on a reviewer agent's feedback systematically — categorize findings, apply fixes, verify tests still pass. Use after request-review returns a report, or when user wants to work through code review findings.
 ---
 

package/run-evals/REFERENCE.md

@@ -0,0 +1,27 @@
+# Run Evals — Reference
+
+## EVALS template
+
+```markdown
+# EVALS: <feature>
+
+## Capability
+| ID | Eval | Grader | verify / rubric |
+|----|------|--------|-----------------|
+| C1 | ... | code | `verify: npm test -- <file>` |
+| C2 | ... | model | Rubric: [ ] criterion A [ ] criterion B |
+
+## Regression
+| ID | Eval | Grader | verify / rubric |
+|----|------|--------|-----------------|
+| R1 | Full suite passes | code | `verify: npm test` |
+
+## Results
+| Run | C1 | C2 | R1 | pass@k |
+|-----|----|----|-----|--------|
+| 1 | PASS | PASS | PASS | 3/3 |
+```
+
+## pass@k
+
+Run capability evals k times (default k=3). Ship when all k pass or document known flake in STATE.md.

package/run-evals/SKILL.md

@@ -0,0 +1,27 @@
+---
+name: run-evals
+description: Eval-Driven Development — define capability and regression evals before building; code graders use verify commands, model graders use explicit rubrics; log pass@k. Use before develop-tdd on new features, or when measuring agent capability over runs.
+model: sonnet
+---
+
+# Run Evals
+
+> **HARD GATE** — Define evals before implementation. Code graders = runnable `verify:` commands; model graders = explicit rubric with pass/fail criteria.
+
+## Process
+
+1. Name the capability under test (one sentence).
+2. Write `specs/EVALS-<feature>.md` with:
+   - **Capability evals** (does it do the job?)
+   - **Regression evals** (did we break anything?)
+3. Assign grader type per eval: `code` (shell verify) or `model` (rubric).
+4. Run evals; log results table with pass@k (e.g. 3/3 runs).
+5. Block BUILD phase until capability evals pass at agreed k.
+
+## Artefact
+
+`specs/EVALS-<feature>.md` — see [REFERENCE.md](REFERENCE.md) for template.
+
+## Verify
+
+→ verify: `ls specs/EVALS-*.md 2>/dev/null | head -1 | grep -q EVALS && echo OK || echo MISSING`

package/scope-work/SKILL.md

@@ -0,0 +1,22 @@
+---
+name: scope-work
+description: Define what is in and out of scope for the current effort and save as specs/SCOPE.md. Use when user wants a PRD, scope definition, or before plan-release on a new initiative.
+model: sonnet
+---
+
+# Scope Work
+
+Turn the current conversation into a bounded PRD at `specs/SCOPE.md`.
+
+## Process
+
+1. Read existing `specs/` artifacts (CONTEXT.md, RELEASE-PLAN.md if any).
+2. Interview (if needed): goal, users, in-scope, out-of-scope, constraints, success metrics.
+3. Write `specs/SCOPE.md` with: Vision, In Scope, Out of Scope, Constraints, Success Criteria.
+4. Run `research-first` if external dependencies are proposed.
+
+> **HARD GATE** — Every "in scope" item must map to a future story ID or explicit "deferred" with reason.
+
+## Verify
+
+→ verify: `test -f specs/SCOPE.md && grep -c "Out of Scope" specs/SCOPE.md | awk '{if($1>0) print "OK"; else print "MISSING"}'`

package/scripts/add-model-frontmatter.sh

@@ -0,0 +1,82 @@
+#!/usr/bin/env bash
+# add-model-frontmatter.sh — one-time helper; idempotent model: injection
+set -euo pipefail
+REPO_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+
+declare -A MODELS=(
+  [using-bigpowers]=sonnet
+  [orchestrate-project]=sonnet
+  [survey-context]=haiku
+  [research-first]=sonnet
+  [elaborate-spec]=opus
+  [map-codebase]=sonnet
+  [model-domain]=sonnet
+  [define-language]=sonnet
+  [grill-me]=sonnet
+  [grill-with-docs]=opus
+  [deepen-architecture]=sonnet
+  [design-interface]=opus
+  [assess-impact]=sonnet
+  [change-request]=sonnet
+  [scope-work]=sonnet
+  [slice-tasks]=sonnet
+  [define-success]=sonnet
+  [plan-work]=opus
+  [plan-refactor]=sonnet
+  [plan-release]=sonnet
+  [spike-prototype]=sonnet
+  [kickoff-branch]=haiku
+  [guard-git]=haiku
+  [hook-commits]=haiku
+  [seed-conventions]=sonnet
+  [develop-tdd]=sonnet
+  [enforce-first]=haiku
+  [delegate-task]=sonnet
+  [dispatch-agents]=sonnet
+  [execute-plan]=haiku
+  [wire-observability]=sonnet
+  [verify-work]=haiku
+  [run-evals]=sonnet
+  [investigate-bug]=sonnet
+  [diagnose-root]=sonnet
+  [validate-fix]=haiku
+  [audit-code]=haiku
+  [request-review]=opus
+  [respond-review]=sonnet
+  [trace-requirement]=haiku
+  [commit-message]=haiku
+  [release-branch]=haiku
+  [inspect-quality]=sonnet
+  [organize-workspace]=haiku
+  [stocktake-skills]=sonnet
+  [evolve-skill]=opus
+  [terse-mode]=haiku
+  [craft-skill]=sonnet
+  [edit-document]=sonnet
+  [session-state]=haiku
+  [migrate-spec]=sonnet
+  [visual-dashboard]=sonnet
+  [write-document]=sonnet
+  [setup-environment]=haiku
+  [reset-baseline]=haiku
+  [search-skills]=haiku
+  [compose-workflow]=sonnet
+  [simulate-agents]=sonnet
+)
+
+for skill_dir in "$REPO_ROOT"/*/; do
+  skill_md="$skill_dir/SKILL.md"
+  [[ -f "$skill_md" ]] || continue
+  name=$(basename "$skill_dir")
+  model="${MODELS[$name]:-sonnet}"
+  if grep -q '^model:' "$skill_md" 2>/dev/null; then
+    continue
+  fi
+  # Insert model: after name: line in frontmatter
+  awk -v m="$model" '
+    /^---$/ { fm++; print; next }
+    fm==1 && /^name:/ { print; print "model: " m; next }
+    { print }
+  ' "$skill_md" > "$skill_md.tmp" && mv "$skill_md.tmp" "$skill_md"
+  echo "model: $model → $name"
+done

package/scripts/build-skill-index.sh

@@ -0,0 +1,28 @@
+#!/usr/bin/env bash
+# build-skill-index.sh — lexical index for search-skills (zero embedding deps)
+set -euo pipefail
+
+REPO_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+OUT="$REPO_ROOT/specs/SKILL-SEARCH-INDEX.md"
+mkdir -p "$(dirname "$OUT")"
+
+{
+  echo "# Skill Search Index (auto-generated)"
+  echo ""
+  echo "Regenerate: \`bash scripts/build-skill-index.sh\`"
+  echo ""
+  echo "| name | model | description |"
+  echo "|------|-------|-------------|"
+  for skill_dir in "$REPO_ROOT"/*/; do
+    skill_md="$skill_dir/SKILL.md"
+    [[ -f "$skill_md" ]] || continue
+    name=$(awk '/^---/{f++} f==1 && /^name:/{print; exit}' "$skill_md" | sed 's/^name:[[:space:]]*//')
+    model=$(awk '/^---/{f++} f==1 && /^model:/{print; exit}' "$skill_md" | sed 's/^model:[[:space:]]*//')
+    desc=$(awk '/^---/{f++} f==1 && /^description:/{p=1} p && !/^---/{print}' "$skill_md" \
+      | sed 's/^description:[[:space:]]*//' | tr -d '\n' | sed 's/|/\\|/g' | cut -c1-200)
+    [[ -z "$name" ]] && continue
+    echo "| $name | ${model:-sonnet} | $desc |"
+  done
+} > "$OUT"
+
+echo "build-skill-index: wrote $OUT"

package/scripts/install.sh

@@ -236,5 +236,9 @@ else
   print_opencode_instructions
   echo ""
   echo "bigpowers installed. Future updates:"
-  echo "  git pull && ./scripts/sync-skills.sh"
+  if [[ -d "$REPO_ROOT/.git" ]]; then
+    echo "  git pull && ./scripts/sync-skills.sh"
+  else
+    echo "  npm update -g bigpowers && bigpowers"
+  fi
 fi

package/scripts/land-branch.sh

@@ -0,0 +1,166 @@
+#!/usr/bin/env bash
+# land-branch.sh — Solo-local integrate: squash-merge feature branch onto main and push.
+# Requires GIT_BIGPOWERS_LAND=1 for hook exceptions on commit/push to protected branches.
+# Usage: bash scripts/land-branch.sh <feature-branch> "<conventional commit message>"
+# Run from the primary repository root (not a linked worktree).
+set -euo pipefail
+
+CONVENTIONAL_REGEX='^(feat|fix|docs|style|refactor|perf|test|build|ci|chore|revert)(\(.+\))?!?: .+'
+
+usage() {
+  echo "Usage: $0 <feature-branch> \"<conventional commit message>\" [--skip-verify]" >&2
+  echo "  Run from primary repo root after release-branch gates (solo-local mode)." >&2
+  exit 1
+}
+
+deny() {
+  echo "ERROR: $1" >&2
+  exit 1
+}
+
+SKIP_VERIFY=false
+ARGS=()
+for arg in "$@"; do
+  if [ "$arg" = "--skip-verify" ]; then
+    SKIP_VERIFY=true
+  else
+    ARGS+=("$arg")
+  fi
+done
+
+FEATURE_BRANCH="${ARGS[0]:-}"
+COMMIT_MSG="${ARGS[1]:-}"
+
+[ -n "$FEATURE_BRANCH" ] && [ -n "$COMMIT_MSG" ] || usage
+
+if [[ ! "$COMMIT_MSG" =~ $CONVENTIONAL_REGEX ]]; then
+  deny "Commit message must follow Conventional Commits: <type>(<scope>): <subject>"
+fi
+
+if [ ${#COMMIT_MSG} -gt 72 ]; then
+  deny "Commit subject line must be 72 characters or less"
+fi
+
+# Primary worktree only (.git is a directory, not a gitdir pointer file)
+if [ -f .git ]; then
+  deny "Run from the primary repository root, not a linked worktree (cd to main repo first)"
+fi
+
+detect_default_branch() {
+  local remote_head
+  remote_head=$(git symbolic-ref refs/remotes/origin/HEAD 2>/dev/null | sed 's@^refs/remotes/origin/@@' || true)
+  if [ -n "$remote_head" ]; then
+    echo "$remote_head"
+    return
+  fi
+  if git show-ref --verify --quiet refs/heads/main; then
+    echo "main"
+  elif git show-ref --verify --quiet refs/heads/master; then
+    echo "master"
+  else
+    deny "Could not detect default branch (main/master)"
+  fi
+}
+
+DEFAULT_BRANCH=$(detect_default_branch)
+REPO_ROOT=$(pwd)
+
+echo "==> Land branch: $FEATURE_BRANCH -> $DEFAULT_BRANCH"
+echo "    Repo root: $REPO_ROOT"
+
+if ! git show-ref --verify --quiet "refs/heads/$FEATURE_BRANCH"; then
+  deny "Feature branch '$FEATURE_BRANCH' does not exist"
+fi
+
+for protected in main master; do
+  if [ "$FEATURE_BRANCH" = "$protected" ]; then
+    deny "Cannot land protected branch '$FEATURE_BRANCH'"
+  fi
+done
+
+run_verify_suite() {
+  echo "==> Running pre-land verification..."
+  if [ -f package.json ] && command -v jq >/dev/null 2>&1; then
+    if jq -e '.scripts.compliance' package.json >/dev/null 2>&1; then
+      npm run compliance
+      return
+    fi
+    if jq -e '.scripts.test' package.json >/dev/null 2>&1; then
+      local test_script
+      test_script=$(jq -r '.scripts.test' package.json)
+      if [ "$test_script" != "echo \"Error: no test specified\" && exit 1" ] && [ "$test_script" != "false" ]; then
+        npm test
+        return
+      fi
+    fi
+    if jq -e '.scripts.lint' package.json >/dev/null 2>&1; then
+      npm run lint
+    fi
+  fi
+  if [ -f scripts/sync-skills.sh ]; then
+    bash scripts/sync-skills.sh
+  fi
+}
+
+if [ "$SKIP_VERIFY" = false ]; then
+  run_verify_suite
+else
+  echo "==> Skipping verification (--skip-verify)"
+fi
+
+echo "==> Updating $DEFAULT_BRANCH"
+git checkout "$DEFAULT_BRANCH"
+if ! git diff-index --quiet HEAD -- 2>/dev/null; then
+  deny "Working tree on $DEFAULT_BRANCH is not clean. Stash or commit first."
+fi
+
+if git remote get-url origin >/dev/null 2>&1; then
+  git pull --ff-only origin "$DEFAULT_BRANCH" || deny "git pull --ff-only failed; resolve before landing"
+fi
+
+if ! git merge-base --is-ancestor "$DEFAULT_BRANCH" "$FEATURE_BRANCH" 2>/dev/null; then
+  deny "Feature branch '$FEATURE_BRANCH' is not based on current $DEFAULT_BRANCH (rebase or recreate branch)"
+fi
+
+export GIT_BIGPOWERS_LAND=1
+
+echo "==> Squash merge $FEATURE_BRANCH"
+git merge --squash "$FEATURE_BRANCH"
+if git diff-index --quiet HEAD -- 2>/dev/null; then
+  deny "Squash merge produced no changes (already merged?)"
+fi
+
+git commit -m "$COMMIT_MSG"
+LAND_SHA=$(git rev-parse --short HEAD)
+echo "==> Land commit: $LAND_SHA"
+
+if git remote get-url origin >/dev/null 2>&1; then
+  echo "==> Pushing $DEFAULT_BRANCH to origin"
+  git push origin "$DEFAULT_BRANCH"
+fi
+
+# Worktree cleanup
+WORKTREE_PATH="../$FEATURE_BRANCH"
+if git worktree list --porcelain 2>/dev/null | grep -q "^worktree $WORKTREE_PATH$"; then
+  echo "==> Removing worktree $WORKTREE_PATH"
+  git worktree remove "$WORKTREE_PATH" 2>/dev/null || git worktree remove -f "$WORKTREE_PATH"
+fi
+git worktree prune 2>/dev/null || true
+
+if git show-ref --verify --quiet "refs/heads/$FEATURE_BRANCH"; then
+  git branch -d "$FEATURE_BRANCH" 2>/dev/null || {
+    echo "WARN: Could not delete branch $FEATURE_BRANCH (not fully merged? use -D manually if intended)"
+  }
+fi
+
+git checkout "$DEFAULT_BRANCH"
+
+echo ""
+echo "Land complete."
+echo "  Branch:   $FEATURE_BRANCH (removed)"
+echo "  Commit:   $LAND_SHA on $DEFAULT_BRANCH"
+echo "  Message:  $COMMIT_MSG"
+echo "  cwd:      $(pwd)"
+echo "  current:  $(git branch --show-current)"
+echo ""
+echo "semantic-release will pick up the push to $DEFAULT_BRANCH when configured."

package/scripts/sync-skills.sh

@@ -102,6 +102,11 @@ jq -n --arg name "bigpowers" \
   echo "}"
 } > "$REPO_ROOT/opencode.json"
 
+# Regenerate lexical skill index for search-skills
+if [[ -x "$REPO_ROOT/scripts/build-skill-index.sh" ]]; then
+  bash "$REPO_ROOT/scripts/build-skill-index.sh" || true
+fi
+
 echo "sync-skills: $skill_count skills synced"
 echo "  → .cursor/rules/ ($skill_count .mdc files)"
 echo "  → .gemini/extensions/bigpowers/skills/ (Agent Skills)"

package/search-skills/SKILL.md

@@ -0,0 +1,20 @@
+---
+name: search-skills
+description: Find the right bigpowers skill from natural-language intent using a local lexical index over SKILL.md frontmatter. Use when unsure which skill to invoke, or at start of research-first.
+model: haiku
+---
+
+# Search Skills
+
+Lexical search only — no embedding service (ADR: zero external dependency).
+
+## Process
+
+1. Run `bash scripts/build-skill-index.sh` if `specs/SKILL-SEARCH-INDEX.md` is stale.
+2. Search index: `rg -i "<keywords>" specs/SKILL-SEARCH-INDEX.md`
+3. Read top 3 matches' `description` and "Use when" triggers.
+4. Recommend one skill with rationale; invoke via orchestrator or direct call.
+
+## Verify
+
+→ verify: `test -f specs/SKILL-SEARCH-INDEX.md && echo OK || (bash scripts/build-skill-index.sh && test -f specs/SKILL-SEARCH-INDEX.md && echo OK)`

package/seed-conventions/SKILL.md

@@ -1,5 +1,6 @@
 ---
 name: seed-conventions
+model: sonnet
 description: Generate CLAUDE.md and CONVENTIONS.md for a brand-new project through a brief interview, and create the specs/ directory. Entry point for greenfield projects. Use when starting a new project from scratch, when user asks to set up AI agent conventions, or when there is no CLAUDE.md yet.
 ---
 
@@ -23,6 +24,8 @@ Ask the user these questions (one at a time, wait for each answer):
 
 2. **Stack** — "What language, framework, and runtime? (e.g. TypeScript / Next.js / Node 22)"
 
+2b. **Stack profile (optional)** — Offer: `swift`, `typescript-vue`, `node-service`, or none. If chosen, merge the matching fragment from `profiles/<name>.md` into generated `CONVENTIONS.md` (commands, architecture, never-do).
+
 3. **Commands** — "What commands do you use for: run, test, build, lint? I'll document them so agents know how to verify their work."
 
 4. **Architecture** — "Describe the key modules and their relationships in 1–2 sentences. What are the main moving parts?"