opencodekit 0.19.4 → 0.19.6

package/dist/index.js

@@ -20,7 +20,7 @@ var __require = /* @__PURE__ */ createRequire(import.meta.url);
 
 //#endregion
 //#region package.json
-var version = "0.19.4";
+var version = "0.19.6";
 
 //#endregion
 //#region src/utils/license.ts

package/dist/template/.opencode/agent/vision.md

@@ -71,7 +71,7 @@ Route by need:
 | Mockup-to-implementation mapping              | `mockup-to-code`      |
 | Distinctive UI direction / anti-slop guidance | `frontend-design`     |
 | Figma design data (read/write via MCP)        | `figma-go`            |
-| Pencil design-as-code workflow                | `pencil`              |
+| OpenPencil design-as-code workflow            | `pencil`              |
 | Brand identity extraction from URLs           | `webclaw`             |
 
 ### Taste-Skill Variants (installed)
@@ -108,12 +108,12 @@ If Figma is available, request MCP access via `figma-go` and ground feedback in
 2. Use `figma-go` to pull `get_design_context` or `get_node`
 3. Reference node IDs in findings for traceability
 
-## Pencil-First Workflow (when no Figma)
+## OpenPencil Workflow (when no Figma)
 
-If design must be created or iterated quickly, use Pencil:
+If design must be created or iterated quickly, use OpenPencil via the legacy `pencil` skill:
 
-1. Create/modify `.pen` via Pencil CLI
-2. Export PNGs for review
+1. Create/modify `.op` via `op` CLI
+2. Export PNGs or code for review
 3. Provide audit with node-level critique where possible
 
 ## Brand Extraction Workflow (when auditing existing sites)

package/dist/template/.opencode/command/compound.md

@@ -1,6 +1,7 @@
 ---
 description: Extract and persist learnings from completed work into institutional memory
 argument-hint: "[bead-id]"
+agent: build
 ---
 
 # Compound: $ARGUMENTS
@@ -19,9 +20,9 @@ so future Plan and Ship cycles start with institutional knowledge instead of bla
 ## Phase 1: Gather Evidence
 
 ```bash
-# Get what changed
-git diff origin/main..HEAD --stat
-git log origin/main..HEAD --oneline
+# Get what changed (falls back gracefully if no remote)
+git diff origin/main..HEAD --stat 2>/dev/null || git diff HEAD~5..HEAD --stat
+git log origin/main..HEAD --oneline 2>/dev/null || git log --oneline -10
 
 # Get review comments if any
 br comments list $ARGUMENTS 2>/dev/null || echo "No bead"

package/dist/template/.opencode/command/create.md

@@ -10,7 +10,7 @@ Create a bead and its specification (PRD) from a description.
 
 > **Workflow:** **`/create`** → `/start <id>` → `/ship <id>`
 >
-> ⛔ This command creates the specification ONLY. Do NOT write any implementation code.
+> This command creates the specification ONLY. Do NOT write any implementation code.
 
 ## Load Skills
 
@@ -133,24 +133,19 @@ Extract bead title and description from `$ARGUMENTS` before creating the bead.
 - If user provided multiple lines, use first line as title and full text as description.
 
 ```bash
-TITLE=$(echo "$ARGUMENTS" | head -n1)
-DESCRIPTION=$(echo "$ARGUMENTS")
-
-if [ "$TITLE" = "$DESCRIPTION" ]; then
-  DESCRIPTION="$TITLE"
-fi
-
 BEAD_ID=$(br create --title "$TITLE" --description "$DESCRIPTION" --type $BEAD_TYPE --json | jq -r '.id')
 mkdir -p ".beads/artifacts/$BEAD_ID"
 ```
 
+Extract title and description from `$ARGUMENTS`:
+- Single line input: use as both title and description
+- Multi-line input: first line as title, full text as description
+
 ## Phase 6: Write PRD
 
 Copy and fill the PRD template using context from Phase 3:
 
-```bash
-cp .opencode/memory/_templates/prd.md ".beads/artifacts/$BEAD_ID/prd.md"
-```
+Read the PRD template from `.opencode/memory/_templates/prd.md` and write it to `.beads/artifacts/$BEAD_ID/prd.md`.
 
 ### Required Sections
 

package/dist/template/.opencode/command/design.md

@@ -32,9 +32,9 @@ skill({ name: "frontend-design" }); // Design system guidance, anti-patterns, re
 ## Phase 1: Detect Existing Design System
 
 ```typescript
-glob({ pattern: "**/tailwind.config.{js,ts,mjs}" });
-glob({ pattern: "**/globals.css" });
-glob({ pattern: "**/components.json" }); // shadcn
+tilth_tilth_files({ pattern: "**/tailwind.config.{js,ts,mjs}" });
+tilth_tilth_files({ pattern: "**/globals.css" });
+tilth_tilth_files({ pattern: "**/components.json" }); // shadcn
 ```
 
 Read what exists. Don't design in a vacuum — build on the project's current system.

package/dist/template/.opencode/command/handoff.md

@@ -33,7 +33,7 @@ br show $ARGUMENTS
 git status --porcelain
 git branch --show-current
 git rev-parse --short HEAD
-ls .beads/artifacts/$ARGUMENTS/ 2>/dev/null
+Read `.beads/artifacts/$ARGUMENTS/` to check existing artifacts.
 ```
 
 ---
@@ -60,10 +60,12 @@ question({
 If user chooses commit:
 
 ```bash
-git add -A
+git add <specific-files-you-modified>
 git commit -m "WIP: $ARGUMENTS - [brief description of where you stopped]"
 ```
 
+**Never use `git add -A` or `git add .`** — stage only the files you modified.
+
 ---
 
 ## Phase 3: Write Handoff

package/dist/template/.opencode/command/init-context.md

@@ -1,14 +1,26 @@
 ---
-description: Initialize GSD-style project planning context with integrated skill usage
+description: Initialize project planning context (roadmap, state) with optional brownfield analysis
 argument-hint: "[--skip-questions] [--brownfield]"
 agent: build
 ---
 
 # Init-Context: $ARGUMENTS
 
-Initialize GSD-style project planning with integrated skill usage.
+Initialize project planning context files from templates.
 
-`tech-stack.md` is the default baseline project memory file. This command adds the optional planning files (`project.md`, `roadmap.md`, `state.md`) used for richer multi-session project context.
+## Architecture
+
+**Auto-injected files** (always loaded into every prompt via `instructions[]`):
+- `user.md` — identity, preferences
+- `tech-stack.md` — framework, constraints
+- `project.md` — vision, success criteria, principles
+- `git-context.md` — spatial awareness
+
+**On-demand files** (created by this command, loaded via `memory-read` when needed):
+- `roadmap.md` — phases, milestones, bead planning
+- `state.md` — current position, blockers, next actions
+
+> **Warning:** Do NOT add roadmap.md or state.md to `instructions[]`. Per-prompt injection of too many files causes session OOM crashes. Use `memory-read({ file: "project/roadmap" })` or `memory-read({ file: "project/state" })` when needed.
 
 ## Load Skills
 
@@ -32,18 +44,20 @@ const args = {
 
 ### 1.1 Check Existing Context
 
-```bash
-ls .opencode/memory/project/ 2>/dev/null && HAS_CONTEXT=true || HAS_CONTEXT=false
-cat .opencode/memory/project/project.md 2>/dev/null | head -20
+Use tilth or Read to check for existing files:
+
+```typescript
+tilth_tilth_files({ pattern: "*.md", scope: ".opencode/memory/project" });
+// Or: Read({ filePath: ".opencode/memory/project/project.md", limit: 20 });
 ```
 
 **If planning context exists:**
 
 ```
-Existing optional planning context found:
-- project.md: [exists/size]
-- roadmap.md: [exists/size]
-- state.md: [exists/size]
+Existing planning context found:
+- project.md: [exists/size] (auto-injected)
+- roadmap.md: [exists/size] (on-demand)
+- state.md: [exists/size] (on-demand)
 
 Options:
 1. Refresh - Delete and recreate from templates
@@ -58,11 +72,10 @@ Wait for user selection.
 If `--brownfield` flag is set:
 
 ```typescript
-// Spawn parallel analysis agents (like GSD map-codebase)
 skill({ name: "swarm-coordination" });
 
 // Agent 1: Map tech stack
-Task({
+task({
   subagent_type: "explore",
   description: "Analyze tech stack",
   prompt:
@@ -70,7 +83,7 @@ Task({
 });
 
 // Agent 2: Map architecture
-Task({
+task({
   subagent_type: "explore",
   description: "Analyze architecture",
   prompt:
@@ -82,12 +95,11 @@ Task({
 
 ## Phase 2: Requirements Gathering
 
-### 2.1 Load Brainstorming Skill (if not --skip-questions)
+### 2.1 Brainstorming (if not --skip-questions)
 
 ```typescript
 if (!args.skipQuestions) {
-  skill({ name: "brainstorming" });
-
+  // brainstorming skill already loaded in Load Skills phase
   // Follow brainstorming process for project vision
   // Ask questions one at a time (as per brainstorming skill)
   // Output: Refined vision, success criteria, target users
@@ -106,12 +118,14 @@ Use template defaults with placeholders for:
 
 ## Phase 3: Document Creation
 
-### 3.1 Create project.md
+### 3.1 Update project.md (auto-injected)
+
+This file is auto-injected into every prompt. Keep it concise.
 
 **Load template:**
 
-```bash
-cat .opencode/memory/_templates/project.md
+```typescript
+Read({ filePath: ".opencode/memory/_templates/project.md" });
 ```
 
 **Fill with gathered data:**
@@ -122,9 +136,15 @@ cat .opencode/memory/_templates/project.md
 - Core principles (convention over config, minimal, extensible)
 - Current phase (from user input or template default)
 
-**Write to:** `.opencode/memory/project/project.md`
+**Write using memory tools:**
+
+```typescript
+memory-update({ file: "project/project", content: filledContent, mode: "replace" });
+```
+
+### 3.2 Create roadmap.md (on-demand)
 
-### 3.2 Create roadmap.md
+This file is NOT auto-injected. Access via `memory-read({ file: "project/roadmap" })`.
 
 **Parse phases from input:**
 
@@ -141,9 +161,15 @@ cat .opencode/memory/_templates/project.md
 | [Phase 1] | [Goal] | [Status] | [#]   |
 ```
 
-**Write to:** `.opencode/memory/project/roadmap.md`
+**Write using memory tools:**
 
-### 3.3 Create state.md
+```typescript
+memory-update({ file: "project/roadmap", content: roadmapContent, mode: "replace" });
+```
+
+### 3.3 Create state.md (on-demand)
+
+This file is NOT auto-injected. Access via `memory-read({ file: "project/state" })`.
 
 **Initialize with:**
 
@@ -157,7 +183,11 @@ cat .opencode/memory/_templates/project.md
 - Open Questions: (empty table)
 - Next Actions: (empty list)
 
-**Write to:** `.opencode/memory/project/state.md`
+**Write using memory tools:**
+
+```typescript
+memory-update({ file: "project/state", content: stateContent, mode: "replace" });
+```
 
 ### 3.4 Brownfield Analysis Integration (if applicable)
 
@@ -166,16 +196,18 @@ If `--brownfield` analysis was run:
 ```typescript
 // Append tech/arch findings to project.md Context Notes section
 // Or create separate .opencode/memory/project/codebase/ documents
-// (similar to GSD's .planning/codebase/ approach)
 ```
 
 ## Phase 4: Verification & Security
 
 ### 4.1 Verify Documents Created
 
-```bash
-ls -la .opencode/memory/project/
-wc -l .opencode/memory/project/*.md
+```typescript
+tilth_tilth_files({ pattern: "*.md", scope: ".opencode/memory/project" });
+// Verify each file exists and has content
+Read({ filePath: ".opencode/memory/project/project.md", limit: 5 });
+Read({ filePath: ".opencode/memory/project/roadmap.md", limit: 5 });
+Read({ filePath: ".opencode/memory/project/state.md", limit: 5 });
 ```
 
 **Check:**
@@ -185,20 +217,19 @@ wc -l .opencode/memory/project/*.md
 - [ ] state.md exists and >20 lines
 - [ ] All files are readable
 
-### 4.2 Secret Scan (Critical - from GSD pattern)
+### 4.2 Secret Scan
 
 ```bash
 # Scan for accidentally leaked secrets in generated docs
-grep -E '(sk-[a-zA-Z0-9]{20,}|sk_live_[a-zA-Z0-9]+|AKIA[A-Z0-9]{16}|ghp_[a-zA-Z0-9]{36}|-----BEGIN.*PRIVATE KEY)' .opencode/memory/project/*.md 2>/dev/null && SECRETS_FOUND=true || SECRETS_FOUND=false
+grep -rE '(sk-[a-zA-Z0-9]{20,}|sk_live_[a-zA-Z0-9]+|AKIA[A-Z0-9]{16}|ghp_[a-zA-Z0-9]{36}|xoxb-[a-zA-Z0-9-]+|-----BEGIN.*PRIVATE KEY)' .opencode/memory/project/*.md 2>/dev/null && SECRETS_FOUND=true || SECRETS_FOUND=false
 ```
 
 **If secrets found:** Alert user and pause before proceeding.
 
-### 4.3 Load Verification Skill
+### 4.3 Run Verification
 
 ```typescript
-skill({ name: "verification-before-completion" });
-
+// verification-before-completion skill already loaded
 // Run verification checklist:
 // 1. IDENTIFY: Files created, structure valid
 // 2. RUN: Validation commands
@@ -207,25 +238,15 @@ skill({ name: "verification-before-completion" });
 // 5. CLAIM: Context initialization complete
 ```
 
-## Phase 5: Beads Integration
-
-### 5.1 Create Initialization Bead (optional)
-
-```bash
-# If user wants to track context setup as a bead
-br create "Initialize project context" --type=task
-br update <bead-id> --status closed --reason="Context files created"
-```
-
 ## Output
 
-Creates optional planning context in `.opencode/memory/project/`:
+Creates planning context in `.opencode/memory/project/`:
 
-| File         | Purpose                                  | Lines (typical) |
-| ------------ | ---------------------------------------- | --------------- |
-| `project.md` | Vision, success criteria, principles     | 50-100          |
-| `roadmap.md` | Phases, milestones, bead planning        | 80-150          |
-| `state.md`   | Current position, blockers, next actions | 60-100          |
+| File         | Purpose                                  | Injection   | Access                                       |
+| ------------ | ---------------------------------------- | ----------- | -------------------------------------------- |
+| `project.md` | Vision, success criteria, principles     | Auto-injected | Updated in-place (already in `instructions[]`) |
+| `roadmap.md` | Phases, milestones, bead planning        | On-demand   | `memory-read({ file: "project/roadmap" })`   |
+| `state.md`   | Current position, blockers, next actions | On-demand   | `memory-read({ file: "project/state" })`     |
 
 **If `--brownfield`:**
 Additional files in `.opencode/memory/project/codebase/`:
@@ -235,21 +256,24 @@ Additional files in `.opencode/memory/project/codebase/`:
 
 ## Success Criteria
 
-- [ ] All `/init-context` planning documents created from templates
+- [ ] All planning documents created from templates
 - [ ] Documents follow template structure
 - [ ] No secrets leaked in generated files
 - [ ] Files pass basic validation (readable, non-empty)
-- [ ] User informed of next steps
+- [ ] User informed of next steps and access patterns
 
-## Phase 6: Custom Context (Optional)
+## Custom Context (Optional)
 
 Inform user about `.opencode/context/` for additional project-specific context:
 
 ```
 Custom context folder available at .opencode/context/
 - Add .md files with architecture decisions, domain knowledge, team agreements
-- Add file paths to opencode.json → instructions[] for AI prompt injection
 - This folder is preserved during init --force and upgrade
+
+⚠️  Only add files to instructions[] if they are essential for EVERY prompt.
+    Per-prompt injection adds ~2-4KB each. Too many files cause session OOM.
+    Prefer memory-read() for on-demand access instead.
 ```
 
 ## Next Steps
@@ -259,7 +283,7 @@ After init-context completes:
 1. **For new projects:** Use `/plan` to create first implementation plan
 2. **For brownfield:** Review codebase analysis, then `/plan`
 3. **For existing beads:** Use `/resume` to continue tracked work
-4. **For custom context:** Add `.md` files to `.opencode/context/` and reference in `opencode.json` instructions
+4. **For custom context:** Add `.md` files to `.opencode/context/` (on-demand via Read, not auto-injected)
 
 ---
 
@@ -267,7 +291,7 @@ After init-context completes:
 
 | Skill                            | When Used                         | Purpose                        |
 | -------------------------------- | --------------------------------- | ------------------------------ |
+| `context-initialization`         | Phase 1                           | Template verification          |
 | `brainstorming`                  | Phase 2 (if not --skip-questions) | Refine vision and requirements |
 | `swarm-coordination`             | Phase 1.2 (if --brownfield)       | Parallel codebase analysis     |
 | `verification-before-completion` | Phase 4                           | Validate created files         |
-| `beads`                          | Phase 5                           | Track as bead if desired       |

package/dist/template/.opencode/command/init-user.md

@@ -83,23 +83,28 @@ updated: [today]
 
 Ensure `user.md` is loaded in `instructions` (bare paths, no `file://` prefix).
 
-`tech-stack.md` is the core project memory file. Other files such as `project.md`, `roadmap.md`, `state.md`, and `gotchas.md` are optional/project-specific and should only be added if they actually exist and you want them injected.
-
-Example minimal setup:
+The default `instructions[]` includes 4 auto-injected files:
 
 ```json
 {
-  "instructions": [".opencode/memory/project/user.md", ".opencode/memory/project/tech-stack.md"]
+  "instructions": [
+    ".opencode/memory/project/user.md",
+    ".opencode/memory/project/tech-stack.md",
+    ".opencode/memory/project/project.md",
+    ".opencode/context/git-context.md"
+  ]
 }
 ```
 
+> **Warning:** Do not add more files to `instructions[]` unless they are essential for every prompt. Per-prompt injection of too many files causes session OOM crashes. Use `memory-read()` for on-demand access instead.
+
 ## Phase 4: Report
 
 Output:
 
 1. user.md created at `.opencode/memory/project/user.md`
 2. Preferences captured
-3. `tech-stack.md` is the only required project memory file by default
-4. Additional project memory files are optional and can be added later
+3. 4 files are auto-injected by default: user.md, tech-stack.md, project.md, git-context.md
+4. Additional planning files (roadmap.md, state.md) are on-demand via `/init-context`
 5. Custom context available at `.opencode/context/` (preserved during init --force and upgrade)
 6. Next step: `/init-context` if the user wants fuller project-planning memory files

package/dist/template/.opencode/command/init.md

@@ -74,9 +74,7 @@ Create `./AGENTS.md` — **target <60 lines** (max 150). Follow the `index-knowl
 
 From template `.opencode/memory/_templates/tech-stack.md`:
 
-```bash
-cp .opencode/memory/_templates/tech-stack.md .opencode/memory/project/tech-stack.md
-```
+Read the template from `.opencode/memory/_templates/tech-stack.md` and write it to `.opencode/memory/project/tech-stack.md`.
 
 Fill detected values:
 

package/dist/template/.opencode/command/lfg.md

@@ -1,6 +1,7 @@
 ---
 description: Full autonomous chain - Plan → Ship → Review → Compound in one command
 argument-hint: "<bead-id> [--skip-plan]"
+agent: build
 ---
 
 # LFG (Let's Fucking Go): $ARGUMENTS
@@ -22,9 +23,10 @@ Full compound engineering cycle. One command, all four steps.
 
 ```bash
 br show $BEAD_ID
-ls .beads/artifacts/$BEAD_ID/
 ```
 
+Read `.beads/artifacts/$BEAD_ID/` to check what artifacts exist.
+
 Verify:
 
 - Bead exists and is `in_progress`
@@ -73,7 +75,7 @@ Checkpoint only at `checkpoint:human-verify` or `checkpoint:decision` tasks.
 ## Step 3: REVIEW
 
 ```bash
-BASE_SHA=$(git rev-parse origin/main 2>/dev/null || git rev-parse HEAD~$(git log --oneline | wc -l | tr -d ' '))
+BASE_SHA=$(git rev-parse origin/main 2>/dev/null || git merge-base HEAD origin/main)
 HEAD_SHA=$(git rev-parse HEAD)
 ```
 

package/dist/template/.opencode/command/plan.md

@@ -10,7 +10,7 @@ Create a detailed implementation plan with TDD steps. Optional deep-planning bet
 
 > **Workflow:** `/create` → `/start <id>` → **`/plan <id>`** (optional) → `/ship <id>`
 >
-> ⛔ Bead MUST be `in_progress` with `prd.md`. Use `/start` first.
+> Bead MUST be `in_progress` with `prd.md`. Use `/start` first.
 >
 > **When to use:** Complex tasks where PRD verification steps aren't enough guidance. Skip for simple tasks.
 
@@ -97,9 +97,10 @@ task({
 
 ```bash
 br show $ARGUMENTS
-ls .beads/artifacts/$ARGUMENTS/
 ```
 
+Read `.beads/artifacts/$ARGUMENTS/` to check what artifacts exist.
+
 Verify:
 
 - Bead is `in_progress`
@@ -125,24 +126,23 @@ Before research, determine discovery level based on PRD:
 **Decision:** Ask user to confirm or adjust:
 
 ```typescript
-const suggestedLevel = assessDiscoveryLevel(prdContent);
-
 question({
   questions: [
     {
       header: "Discovery Level",
-      question: `Suggested: Level ${suggestedLevel} (${getLevelDescription(suggestedLevel)}). Proceed?`,
+      question: "Suggested discovery level based on PRD complexity. Proceed?",
       options: [
-        { label: `Yes, Level ${suggestedLevel} (Recommended)` },
-        { label: "Lower (less research)", description: "If you know the patterns" },
-        { label: "Higher (more research)", description: "If uncertain about approach" },
-        { label: "Skip research", description: "I know the codebase" },
+        { label: "Deep (Recommended for complex work)", description: "Level 2-3: spawn scout + explore agents" },
+        { label: "Standard", description: "Level 1: quick doc lookup" },
+        { label: "Skip research", description: "Level 0: I know the codebase" },
       ],
     },
   ],
 });
 ```
 
+Determine level from PRD content: Level 2+ if new library, external API, or "choose/evaluate" language. Level 3 if "architecture/design/system".
+
 ## Phase 3: Research (if Level 1-3)
 
 Read the PRD and extract tasks, success criteria, affected files, scope.

package/dist/template/.opencode/command/pr.md

@@ -53,9 +53,10 @@ If bead ID provided:
 
 ```bash
 br show $ARGUMENTS
-ls .beads/artifacts/$ARGUMENTS/
 ```
 
+Read `.beads/artifacts/$ARGUMENTS/` to check what artifacts exist.
+
 Read the PRD to extract goal and success criteria for the PR description.
 
 ## Phase 2B: Pre-PR Review
@@ -115,6 +116,9 @@ git push -u origin $(git branch --show-current)
 ## Phase 4: Create PR
 
 ```bash
+# Verify gh CLI is installed
+command -v gh >/dev/null 2>&1 || { echo "Error: gh CLI not found. Install: https://cli.github.com"; exit 1; }
+
 gh pr create --title "<title>" --body "$(cat <<'EOF'
 ## Summary
 

package/dist/template/.opencode/command/resume.md

@@ -33,13 +33,13 @@ If not on the right branch, check out the feature branch. If uncommitted changes
 
 ## Phase 3: Find Handoff
 
-Check for handoff notes:
+Check for handoff notes in the memory system:
 
-```bash
-ls .beads/artifacts/$ARGUMENTS/handoffs/ 2>/dev/null
+```typescript
+memory_read({ file: "handoffs/$ARGUMENTS" });
 ```
 
-If a handoff exists, read the latest one. It tells you:
+If a handoff exists, it tells you:
 
 - What was completed
 - Where work stopped

package/dist/template/.opencode/command/review-codebase.md

@@ -37,7 +37,7 @@ skill({ name: "requesting-code-review" });
 | `explore`    | Finding patterns in codebase, prior art |
 | `scout`      | External research, best practices       |
 | `lsp`        | Finding symbol definitions, references  |
-| `grep`       | Finding code patterns                   |
+| `tilth_tilth_search` | Finding code patterns |
 | `codesearch` | Real-world usage examples               |
 
 ## Phase 1: Gather Context
@@ -66,7 +66,7 @@ If bead provided, read `.beads/artifacts/$ID/prd.md` to review against spec.
 
 If bead provided, read `.beads/artifacts/$ID/prd.md` to review against spec.
 
-## Phase 2: Automated Checks
+## Phase 3: Automated Checks
 
 Detect project type and run the appropriate checks in parallel:
 
@@ -86,7 +86,7 @@ Also scan for common issues appropriate to the detected language:
 - `TODO|FIXME|HACK` markers
 - Hardcoded secrets patterns
 
-## Phase 3: Manual Review
+## Phase 4: Manual Review
 
 Review each category:
 
@@ -105,7 +105,7 @@ Review each category:
 - Default: Full automated + manual review
 - `--thorough`: Deep analysis of all categories
 
-## Phase 4: Report
+## Phase 5: Report
 
 Group findings by severity: