specweave 1.0.437 → 1.0.439

package/plugins/specweave/skills/pr/SKILL.md

@@ -0,0 +1,215 @@
+---
+description: Create pull request from increment feature branch. Use when increment is complete and push strategy is pr-based, or when explicitly saying "create PR", "open pull request", "make a PR".
+argument-hint: "<increment-id>"
+user-invokable: true
+---
+
+# Create Pull Request
+
+Creates a pull request for a completed increment. Automatically invoked by `sw:done` when `cicd.pushStrategy` is `"pr-based"`. Can also be invoked manually.
+
+## When to Activate
+
+**Do activate:**
+- `sw:done` invokes this after quality gates pass (when `pushStrategy: "pr-based"`)
+- User says "create PR", "open pull request", "make a PR for this increment"
+- User says "push this as a PR"
+
+**Do NOT activate:**
+- `pushStrategy` is `"direct"` and user didn't explicitly ask for a PR
+- Increment is not active or completed
+- No changes exist to push
+
+## Step 1: Read Configuration
+
+```bash
+# Read push strategy and git config
+PUSH_STRATEGY=$(jq -r '.cicd.pushStrategy // "direct"' .specweave/config.json 2>/dev/null)
+BRANCH_PREFIX=$(jq -r '.cicd.git.branchPrefix // "sw/"' .specweave/config.json 2>/dev/null)
+TARGET_BRANCH=$(jq -r '.cicd.git.targetBranch // "main"' .specweave/config.json 2>/dev/null)
+DELETE_ON_MERGE=$(jq -r '.cicd.git.deleteOnMerge // true' .specweave/config.json 2>/dev/null)
+```
+
+Read increment metadata to check if PR already exists:
+```bash
+jq -r '.prRefs // empty' .specweave/increments/{increment-id}/metadata.json
+```
+
+If `prRefs` already has an entry with `state: "open"`, skip PR creation and report existing PR URL.
+
+## Step 2: Determine Branch Name
+
+```bash
+BRANCH_NAME="${BRANCH_PREFIX}${INCREMENT_ID}"
+# e.g., sw/0520-pr-based-increment-closure
+```
+
+Check current branch:
+```bash
+CURRENT_BRANCH=$(git rev-parse --abbrev-ref HEAD)
+```
+
+## Step 3: Ensure on Feature Branch
+
+**If already on the feature branch** (`CURRENT_BRANCH === BRANCH_NAME`):
+- Skip branch creation, proceed to push.
+
+**If on target branch** (main/develop):
+- Check if feature branch exists locally: `git branch --list ${BRANCH_NAME}`
+- If exists: `git checkout ${BRANCH_NAME}`
+- If not: `git checkout -b ${BRANCH_NAME}`
+
+**If on a different branch** (user created manually):
+- Use the existing branch as-is. Set `BRANCH_NAME = CURRENT_BRANCH`.
+- Do NOT force rename to match naming convention.
+
+## Step 4: Push Branch
+
+```bash
+git push -u origin ${BRANCH_NAME}
+```
+
+If push fails (e.g., no remote, auth issues), warn and exit gracefully. Do NOT block increment closure.
+
+## Step 5: Build PR Description
+
+Read spec.md to build the PR body. Use `--body-file` to avoid shell quoting issues:
+
+```bash
+# Create temp file for PR body
+cat > /tmp/sw-pr-body-${INCREMENT_ID}.md << 'PREOF'
+## Summary
+
+<!-- Auto-generated from SpecWeave increment {INCREMENT_ID} -->
+
+{spec.md overview section — first paragraph}
+
+## User Stories
+
+{For each US in spec.md:}
+- **US-001**: {title} ({AC count} acceptance criteria)
+- **US-002**: {title} ({AC count} acceptance criteria)
+
+## Acceptance Criteria
+
+{Checklist of all ACs from spec.md:}
+- [ ] AC-US1-01: {criterion}
+- [ ] AC-US1-02: {criterion}
+
+## Test Summary
+
+{Summary of test plan from tasks.md — what was tested, coverage}
+
+---
+
+*Created by [SpecWeave](https://spec-weave.com) increment `{INCREMENT_ID}`*
+PREOF
+```
+
+## Step 6: Create Pull Request
+
+```bash
+gh pr create \
+  --title "[${INCREMENT_ID}] {spec title}" \
+  --body-file /tmp/sw-pr-body-${INCREMENT_ID}.md \
+  --base ${TARGET_BRANCH} \
+  --head ${BRANCH_NAME}
+```
+
+Capture the PR URL and number:
+```bash
+PR_URL=$(gh pr create ... 2>&1)
+PR_NUMBER=$(echo "$PR_URL" | grep -oE '[0-9]+$')
+```
+
+Clean up temp file:
+```bash
+rm -f /tmp/sw-pr-body-${INCREMENT_ID}.md
+```
+
+If `deleteOnMerge` is true, add auto-delete label or note in PR description.
+
+## Step 7: Update Metadata
+
+Update increment metadata with PR reference. Edit `metadata.json` to add/update `prRefs`:
+
+```json
+{
+  "prRefs": [{
+    "branch": "sw/0520-pr-based-increment-closure",
+    "prNumber": 42,
+    "prUrl": "https://github.com/org/repo/pull/42",
+    "state": "open",
+    "createdAt": "2026-03-12T10:00:00Z"
+  }]
+}
+```
+
+Use `jq` or direct JSON edit to update metadata.json.
+
+## Step 8: Multi-Repo / Umbrella Mode
+
+Check if umbrella mode is enabled:
+```bash
+UMBRELLA=$(jq -r '.umbrella.enabled // false' .specweave/config.json 2>/dev/null)
+```
+
+If umbrella mode:
+
+1. **Scan for modified repos**: Check each `repositories/{org}/{repo}/` for changes relative to target branch:
+   ```bash
+   for repo_dir in repositories/*/*; do
+     if [ -d "$repo_dir/.git" ]; then
+       cd "$repo_dir"
+       CHANGES=$(git log ${TARGET_BRANCH}..HEAD --oneline 2>/dev/null | wc -l)
+       if [ "$CHANGES" -gt 0 ]; then
+         # This repo has changes — create branch + PR
+       fi
+       cd -
+     fi
+   done
+   ```
+
+2. **Create branch + PR in each repo** with changes (Steps 3-6 per repo).
+
+3. **Collect all PR refs** and store in metadata as array:
+   ```json
+   {
+     "prRefs": [
+       { "branch": "sw/0520-...", "prUrl": "...repo1/pull/42", "repoSlug": "org/repo1" },
+       { "branch": "sw/0520-...", "prUrl": "...repo2/pull/15", "repoSlug": "org/repo2" }
+     ]
+   }
+   ```
+
+4. **Continue on failure**: If one repo's PR fails, log the error and continue with others. Report all results at the end.
+
+## Step 9: Output Summary
+
+Display a summary table:
+
+```
+PR CREATION SUMMARY
+═══════════════════════════════════════════════
+Increment: 0520-pr-based-increment-closure
+Strategy:  pr-based
+Target:    main
+
+  Repo            Branch                          PR
+  ─────────────   ─────────────────────────────   ────────────────────────────
+  org/repo1       sw/0520-pr-based-closure        https://github.com/.../42
+  org/repo2       sw/0520-pr-based-closure        https://github.com/.../15
+
+Status: 2/2 PRs created successfully
+```
+
+## Error Handling
+
+| Error | Action |
+|-------|--------|
+| `gh` not installed | Warn: "Install GitHub CLI: brew install gh" |
+| `gh auth` not configured | Warn: "Run: gh auth login" |
+| Push rejected | Warn but don't block. Show error message. |
+| PR already exists for branch | Skip creation, report existing PR URL |
+| No changes to push | Skip PR creation, inform user |
+| Merge conflicts with target | Create PR anyway (conflicts visible in PR UI) |

package/plugins/specweave/skills/team-build/SKILL.md

@@ -157,15 +157,15 @@ This spawns three parallel reviewers:
 
 Generate comprehensive test coverage across all test levels simultaneously. Each agent focuses on a different testing layer and operates independently.
 
-> **Note:** `testing:qa` is the primary orchestration skill for testing workflows. This preset splits its responsibilities into specialized agents for parallel execution.
+> **Note:** SpecWeave testing skills (`sw:tdd-red`, `sw:e2e`, `sw:validate`) provide the testing workflows. This preset splits responsibilities into specialized agents for parallel execution.
 
 #### Agent Composition
 
 | # | Role | Skill(s) | Owns | Responsibility |
 |---|------|----------|------|----------------|
-| 1 | Unit | `testing:unit` | `tests/unit/` | Write unit tests for individual functions, classes, and modules with proper mocking |
-| 2 | E2E | `testing:e2e` | `tests/e2e/` | Write end-to-end tests for user flows, API sequences, and cross-service interactions |
-| 3 | Coverage | `testing:qa` | `tests/` (analysis scope) | Analyze coverage gaps, generate missing test cases, ensure threshold compliance |
+| 1 | Unit | `sw:tdd-red` | `tests/unit/` | Write unit tests for individual functions, classes, and modules with proper mocking |
+| 2 | E2E | `sw:e2e` | `tests/e2e/` | Write end-to-end tests for user flows, API sequences, and cross-service interactions |
+| 3 | Coverage | `sw:validate` | `tests/` (analysis scope) | Analyze coverage gaps, generate missing test cases, ensure threshold compliance |
 
 #### Execution Chain
 

package/plugins/specweave/skills/team-lead/SKILL.md

@@ -164,7 +164,7 @@ Analyze the feature request and map affected domains to SpecWeave skills.
 | **Backend** | `sw:architect` | `infra:devops` | API endpoints, services, business logic |
 | **Database** | `sw:architect` | | Schema design, migrations, seed data |
 | **Shared/Types** | `sw:architect` | `sw:code-simplifier` | TypeScript interfaces, shared constants, API contracts |
-| **Testing** | `testing:qa` | `testing:e2e`, `testing:unit` | Test strategy, E2E suites, integration tests |
+| **Testing** | `sw:e2e` | `sw:tdd-red`, `sw:validate` | Test strategy, E2E suites, integration tests |
 | **Security** | `sw:security` | `security:patterns` | Auth, authorization, threat modeling, OWASP |
 | **DevOps** | `infra:devops` | `k8s:deployment-generate`, `infra:observability` | CI/CD, Docker, K8s, monitoring |
 | **Mobile** | `mobile:react-native` | `mobile:screen-generate`, `mobile:expo` | Native/cross-platform mobile apps |
@@ -416,7 +416,7 @@ Agent definitions live as reusable `.md` files in the `agents/` subdirectory. Wh
 | Frontend | `agents/frontend.md` | UI, components, pages | 2 (downstream) | `frontend:architect`, `frontend:design` |
 | Backend | `agents/backend.md` | API, services, middleware | 2 (downstream) | `sw:architect`, `infra:devops` |
 | Database | `agents/database.md` | Schema, migrations, seeds | 1 (upstream) | `sw:architect` |
-| Testing | `agents/testing.md` | Unit, integration, E2E | 2 (downstream) | `testing:qa`, `testing:e2e` |
+| Testing | `agents/testing.md` | Unit, integration, E2E | 2 (downstream) | `sw:e2e`, `sw:tdd-red` |
 | Security | `agents/security.md` | Auth, validation, audit | 2 (downstream) | `sw:security` |
 
 ### How to Use Agent Files

package/plugins/specweave/skills/team-lead/agents/testing.md

@@ -7,9 +7,9 @@ MASTER SPEC (SOURCE OF TRUTH):
   Read the master spec BEFORE planning any work.
 
 SKILLS TO INVOKE:
-  Skill({ skill: "testing:qa" })
-  Skill({ skill: "testing:e2e" })        // for E2E test suites
-  Skill({ skill: "testing:unit" })       // for unit test coverage
+  Skill({ skill: "sw:e2e", args: "--generate [INCREMENT_ID]" })  // generate E2E tests from ACs
+  Skill({ skill: "sw:e2e", args: "--run [INCREMENT_ID]" })       // run E2E + produce e2e-report.json
+  Skill({ skill: "sw:e2e", args: "--a11y [INCREMENT_ID]" })      // E2E + accessibility audit
 
 FILE OWNERSHIP (WRITE access):
   tests/**

package/plugins/specweave-github/skills/pr-review/SKILL.md

@@ -0,0 +1,166 @@
+---
+description: AI-powered pull request review against spec acceptance criteria. Use for "review PR", "check PR against spec", "review pull request". Enterprise feature.
+argument-hint: "<increment-id|pr-url>"
+user-invokable: true
+---
+
+# AI Pull Request Review
+
+Reviews a pull request against the increment's spec.md acceptance criteria. Posts structured review comments on the PR via `gh pr review`.
+
+This is an **enterprise feature** — optional, not part of the default flow. Invoke explicitly or configure for automatic invocation.
+
+## When to Activate
+
+**Do activate:**
+- User says "review PR", "check PR against spec", "review pull request"
+- User says "AI review for PR #42"
+- Configured for automatic review after `sw:pr` creates a PR
+
+**Do NOT activate:**
+- User is doing manual code review (don't interfere)
+- No increment context available
+- PR is already merged
+
+## Step 1: Resolve PR and Increment
+
+**If given an increment ID:**
+```bash
+# Read PR refs from metadata
+PR_URL=$(jq -r '.prRefs[0].prUrl // empty' .specweave/increments/{id}/metadata.json)
+PR_NUMBER=$(jq -r '.prRefs[0].prNumber // empty' .specweave/increments/{id}/metadata.json)
+```
+
+**If given a PR URL or number:**
+```bash
+# Extract PR number from URL
+PR_NUMBER=$(echo "$PR_URL" | grep -oE '[0-9]+$')
+# Find increment by searching metadata for matching prRefs
+```
+
+**If neither:** Check current branch for an associated PR:
+```bash
+CURRENT_BRANCH=$(git rev-parse --abbrev-ref HEAD)
+PR_NUMBER=$(gh pr list --head "$CURRENT_BRANCH" --json number -q '.[0].number')
+```
+
+## Step 2: Load Spec Context
+
+Read the increment spec.md to extract:
+1. All acceptance criteria (AC-USXX-YY)
+2. User stories overview
+3. Non-functional requirements
+4. Edge cases documented in the spec
+
+Build a review checklist from ACs.
+
+## Step 3: Get PR Diff
+
+```bash
+gh pr diff ${PR_NUMBER} > /tmp/sw-pr-diff-${PR_NUMBER}.diff
+```
+
+Also get the list of changed files:
+```bash
+gh pr view ${PR_NUMBER} --json files -q '.files[].path'
+```
+
+## Step 4: Review Against Acceptance Criteria
+
+For each AC in spec.md, analyze the diff to determine:
+
+- **SATISFIED**: The diff clearly implements this criterion
+- **PARTIALLY SATISFIED**: Some aspects are covered, others are missing
+- **NOT SATISFIED**: No evidence of implementation in the diff
+- **NOT APPLICABLE**: This AC is not relevant to the changed files
+
+Build a structured review:
+
+```markdown
+## SpecWeave AC Review — Increment {INCREMENT_ID}
+
+### Acceptance Criteria Coverage
+
+| AC | Status | Evidence |
+|----|--------|----------|
+| AC-US1-01: User can log in | SATISFIED | `src/auth/login.ts` implements full flow |
+| AC-US1-02: Invalid creds show error | PARTIALLY | Error handling exists but no user-facing message |
+| AC-US2-01: Session persists | NOT SATISFIED | No session storage implementation found |
+
+### Code Quality Observations
+
+{List any code quality issues found in the diff — not bugs, but patterns that deviate from the spec or best practices}
+
+### Summary
+
+{Overall assessment: ready to merge / needs changes / needs discussion}
+- {X}/{Y} ACs satisfied
+- {Z} observations
+```
+
+## Step 5: Post Review
+
+Determine review action based on AC coverage:
+
+- **All ACs satisfied**: `--approve`
+- **Any AC not satisfied**: `--request-changes`
+- **Only observations, no blockers**: `--comment`
+
+```bash
+gh pr review ${PR_NUMBER} \
+  --body-file /tmp/sw-pr-review-${PR_NUMBER}.md \
+  --{approve|request-changes|comment}
+```
+
+Clean up:
+```bash
+rm -f /tmp/sw-pr-diff-${PR_NUMBER}.diff /tmp/sw-pr-review-${PR_NUMBER}.md
+```
+
+## Step 6: Post Inline Comments (Optional)
+
+For specific issues found in the diff, post inline comments on the relevant lines:
+
+```bash
+# For each finding with a specific file + line:
+gh api repos/{owner}/{repo}/pulls/${PR_NUMBER}/comments \
+  -f body="**AC-US1-02**: Error message is logged but not displayed to the user. The spec requires a user-facing error toast." \
+  -f commit_id="{latest_commit_sha}" \
+  -f path="src/auth/login.ts" \
+  -F line=42 \
+  -f side="RIGHT"
+```
+
+## Scheduled Review Mode
+
+For teams that want automated PR reviews, this skill can be triggered via Claude Code scheduled tasks:
+
+```json
+// .claude/scheduled-tasks.json
+{
+  "tasks": [{
+    "schedule": "*/30 * * * *",
+    "command": "/sw:pr-review --all-open",
+    "description": "Review all open PRs every 30 minutes"
+  }]
+}
+```
+
+When invoked with `--all-open`:
+1. List all open PRs: `gh pr list --state open --json number,headRefName`
+2. For each PR, check if it has an associated increment (via branch naming convention or metadata scan)
+3. Skip PRs already reviewed by this tool (check for existing SpecWeave review comments)
+4. Review each PR against its increment spec
+
+## Multi-Agent Review (Complex PRs)
+
+For PRs spanning multiple domains (frontend + backend + infrastructure), consider using `sw:team-build` with the `review` preset to spawn domain-specialized review agents in parallel.
+
+## Error Handling
+
+| Error | Action |
+|-------|--------|
+| No spec.md found | Warn: "No increment spec found for this PR. Cannot perform AC review." |
+| PR already merged | Skip: "PR #{N} is already merged." |
+| `gh` auth issues | Warn: "Run: gh auth login" |
+| Large diff (>5000 lines) | Summarize by file instead of line-by-line review |