@enderfga/claw-orchestrator 3.0.0

package/skills/references/ultra.md

@@ -0,0 +1,126 @@
+# Ultraplan & Ultrareview
+
+High-level orchestration features for deep project planning and fleet code review. Inspired by Claude Code's Ultraplan and Ultrareview commands.
+
+## Ultraplan
+
+A dedicated planning session that explores your project thoroughly (up to 30 minutes) and produces a detailed implementation plan.
+
+### How It Works
+
+1. Starts an Opus session in **plan mode** with **max effort**
+2. Appends a system prompt instructing thorough exploration and plan-only output
+3. Sends the task description and waits for completion
+4. Returns the plan text when done
+
+Runs in background — poll with `ultraplan_status`.
+
+### Usage
+
+```typescript
+const plan = manager.ultraplanStart('Add OAuth2 support with Google and GitHub providers', {
+  cwd: '/path/to/project',
+  model: 'opus',           // default
+  timeout: 1800000,        // 30 min default
+});
+
+console.log(`Plan ID: ${plan.id}`);
+
+// Poll for completion
+const status = manager.ultraplanStatus(plan.id);
+if (status?.status === 'completed') {
+  console.log(status.plan);
+}
+```
+
+### Tools
+
+| Tool | Description |
+|------|-------------|
+| `ultraplan_start` | Start planning session (background) |
+| `ultraplan_status` | Get status and plan text |
+
+### Configuration
+
+| Parameter | Default | Description |
+|-----------|---------|-------------|
+| `model` | `opus` | Model to use for planning |
+| `cwd` | `process.cwd()` | Project directory to explore |
+| `timeout` | 1,800,000 ms (30 min) | Maximum planning time |
+
+Results remain queryable for 30 minutes after completion.
+
+---
+
+## Ultrareview
+
+A fleet of specialized bug-hunting agents that review your codebase in parallel, each from a different angle. Built on top of the [Council](./council.md) system.
+
+### How It Works
+
+1. Creates a council with N reviewer agents (5-20)
+2. Each agent specializes in a different review angle
+3. Agents run in parallel via git worktree isolation
+4. Findings from all agents are synthesized into a single report
+
+### Available Review Angles (20)
+
+| Agent | Focus |
+|-------|-------|
+| SecurityReviewer | Injection, auth flaws, data exposure, OWASP top 10 |
+| LogicReviewer | Off-by-one, race conditions, null handling, edge cases |
+| PerformanceReviewer | O(n^2) loops, memory leaks, missing caching, N+1 queries |
+| APIReviewer | Inconsistent interfaces, missing validation, error gaps |
+| TestReviewer | Untested paths, missing edge case tests, flaky patterns |
+| TypeReviewer | `any` casts, unsafe assertions, missing null checks |
+| ConcurrencyReviewer | Race conditions, deadlocks, async error handling |
+| ErrorReviewer | Swallowed errors, missing try/catch, crash paths |
+| DependencyReviewer | Outdated packages, CVEs, unnecessary deps |
+| ReadabilityReviewer | Unclear naming, complex functions, dead code |
+| DataReviewer | Data validation, schema mismatches, encoding |
+| ConfigReviewer | Hardcoded values, missing env vars, insecure defaults |
+| ScalabilityReviewer | Single points of failure, unbounded growth |
+| DocReviewer | Outdated docs, missing API docs, misleading comments |
+| A11yReviewer | ARIA labels, keyboard nav, color contrast |
+| I18nReviewer | Hardcoded strings, locale handling, RTL support |
+| NetworkReviewer | Missing timeouts, retry logic, connection pooling |
+| AuthReviewer | Token handling, CSRF, permission checks |
+| CryptoReviewer | Weak algorithms, key management, RNG |
+| MemoryReviewer | Memory leaks, circular references, stream handling |
+
+### Usage
+
+```typescript
+const review = manager.ultrareviewStart('/path/to/project', {
+  agentCount: 10,            // use 10 of the 20 angles
+  maxDurationMinutes: 15,    // 15 min timeout per agent
+  model: 'sonnet',           // model for all reviewers
+  focus: 'Find security and performance bugs',
+});
+
+console.log(`Review ID: ${review.id}, Council: ${review.councilId}`);
+
+// Poll for completion
+const status = manager.ultrareviewStatus(review.id);
+if (status?.status === 'completed') {
+  console.log(status.findings);
+}
+```
+
+### Tools
+
+| Tool | Description |
+|------|-------------|
+| `ultrareview_start` | Launch reviewer fleet (background) |
+| `ultrareview_status` | Get status and findings |
+
+### Configuration
+
+| Parameter | Default | Range | Description |
+|-----------|---------|-------|-------------|
+| `agentCount` | 5 | 1-20 | Number of reviewer agents |
+| `maxDurationMinutes` | 10 | 5-25 | Per-agent timeout |
+| `model` | session default | — | Model for all reviewers |
+| `focus` | bugs + security + quality | — | Review focus description |
+
+The council runs with `maxRounds: 2` — one round to find bugs, one to cross-review. Results remain queryable for 30 minutes.