milens 0.7.0 → 0.7.2

package/.agents/skills/adapters/SKILL.md

@@ -1,51 +1,59 @@
----
-name: milens-adapters
-description: Code intelligence for the adapters area — symbols, dependencies, and entry points
----
-
-# Adapters
-
-## Working with this area
-When working with code in **adapters/**, follow these mandatory safety rules:
-
-### Before editing any symbol in this area:
-1. Call `mcp_milens_impact({target: "<symbol>", repo: "<workspaceRoot>"})` — check blast radius
-2. If depth-1 dependents > 5 → **STOP and warn** before proceeding
-3. Call `mcp_milens_context({name: "<symbol>", repo: "<workspaceRoot>"})` — see all callers/callees
-
-### Before committing changes in this area:
-1. Call `mcp_milens_detect_changes({repo: "<workspaceRoot>"})` — verify scope
-2. If unexpected files changed → **STOP and report**
-
-### Key tools for this area:
-| Task | Tool |
-|---|---|
-| Find all references | `mcp_milens_context` |
-| Check edit safety | `mcp_milens_edit_check` |
-| Text search across files | `mcp_milens_grep` |
-| See file symbols | `mcp_milens_get_file_symbols` |
-
-## Overview
-Contains 34 symbols (31 exported) across 3 files.
-
-## Key Symbols
-- **`Milens — Code Intelligence for Claude Code`** [section] (adapters/claude-code/CLAUDE.md:1) — 0 refs
-- **`⚠️ BEFORE ANYTHING ELSE`** [section] (adapters/claude-code/CLAUDE.md:3) — 0 refs
-- **`⭐ Core Tools (Use Every Session)`** [section] (adapters/claude-code/CLAUDE.md:19) — 0 refs
-- **`🔧 Situational Tools (Use When Needed)`** [section] (adapters/claude-code/CLAUDE.md:32) — 0 refs
-- **`Session Workflow`** [section] (adapters/claude-code/CLAUDE.md:46) — 0 refs
-- **`Edit Safety (Mandatory)`** [section] (adapters/claude-code/CLAUDE.md:54) — 0 refs
-- **`Tool Selection`** [section] (adapters/claude-code/CLAUDE.md:70) — 0 refs
-- **`Reference`** [section] (adapters/claude-code/CLAUDE.md:76) — 0 refs
-- **`Milens — Code Intelligence for OpenCode`** [section] (adapters/opencode/AGENTS.md:1) — 0 refs
-- **`⚠️ BEFORE ANYTHING ELSE`** [section] (adapters/opencode/AGENTS.md:3) — 0 refs
-- **`⭐ Core Tools (Use Every Session)`** [section] (adapters/opencode/AGENTS.md:19) — 0 refs
-- **`🔧 Situational Tools (Use When Needed)`** [section] (adapters/opencode/AGENTS.md:32) — 0 refs
-- **`Session Workflow`** [section] (adapters/opencode/AGENTS.md:46) — 0 refs
-- **`Edit Safety (Mandatory)`** [section] (adapters/opencode/AGENTS.md:54) — 0 refs
-- **`Tool Selection`** [section] (adapters/opencode/AGENTS.md:70) — 0 refs
-
-## Files
-- adapters/README.md
-- adapters/claude-code/CLAUDE.md
-- adapters/opencode/AGENTS.md
+---
+generated-by: milens
+name: milens-adapters
+description: Code intelligence for the adapters area — symbols, dependencies, and entry points
+---
+
+<!-- Generated by milens — safe to delete/regenerate -->
+# Adapters
+
+## Working with this area
+When working with code in **adapters/**, follow these mandatory safety rules:
+
+### Before editing any symbol in this area:
+1. Call `mcp_milens_impact({target: "<symbol>", repo: "<workspaceRoot>"})` — check blast radius
+2. If depth-1 dependents > 5 → **STOP and warn** before proceeding
+3. Call `mcp_milens_context({name: "<symbol>", repo: "<workspaceRoot>"})` — see all callers/callees
+
+### Before committing changes in this area:
+1. Call `mcp_milens_detect_changes({repo: "<workspaceRoot>"})` — verify scope
+2. If unexpected files changed → **STOP and report**
+
+### Key tools for this area:
+| Task | Tool |
+|---|---|
+| Find all references | `mcp_milens_context` |
+| Check edit safety | `mcp_milens_edit_check` |
+| Text search across files | `mcp_milens_grep` |
+| See file symbols | `mcp_milens_get_file_symbols` |
+
+### Edit-safety enforcement
+A `PreToolUse` hook (warn mode by default) reminds you if no milens safety check (`impact`/`context`/`overview`/`guard_edit_check`/`edit_check`/`smart_context`) was called before an `Edit`/`Write`/`MultiEdit`. Opt-in strict deny mode is available via `milens hooks guard-set-mode --mode strict`. Both modes consume the check after one edit. See `.milens/hook-state/config.json`. Known caveat: the underlying `PreToolUse` deny mechanism has at least one reliability issue (https://github.com/anthropics/claude-code/issues/4362).
+
+## Overview
+Contains 61 symbols (55 exported) across 6 files.
+
+## Key Symbols
+- **`Milens Adapter Packs`** [section] (adapters/README.md:1) — 0 refs
+- **`What is milens?`** [section] (adapters/README.md:5) — 0 refs
+- **`Available Adapters`** [section] (adapters/README.md:9) — 0 refs
+- **`Installation`** [section] (adapters/README.md:21) — 0 refs
+- **`Claude Code`** [section] (adapters/README.md:23) — 0 refs
+- **`Claude Code — via `/plugin install` (marketplace)`** [section] (adapters/README.md:50) — 0 refs
+- **`OpenCode`** [section] (adapters/README.md:65) — 0 refs
+- **`Codex`** [section] (adapters/README.md:72) — 0 refs
+- **`Cursor`** [section] (adapters/README.md:78) — 0 refs
+- **`GitHub Copilot`** [section] (adapters/README.md:85) — 0 refs
+- **`Gemini`** [section] (adapters/README.md:92) — 0 refs
+- **`Zed`** [section] (adapters/README.md:98) — 0 refs
+- **`Verification`** [section] (adapters/README.md:104) — 0 refs
+- **`Index Maintenance`** [section] (adapters/README.md:112) — 0 refs
+- **`All Tools Available`** [section] (adapters/README.md:120) — 0 refs
+
+## Files
+- adapters/README.md
+- adapters/claude-code/CLAUDE.md
+- adapters/codex/.codex/codex.md
+- adapters/copilot/.github/copilot-instructions.md
+- adapters/gemini/.gemini/context.md
+- adapters/opencode/AGENTS.md

package/.agents/skills/analyzer/SKILL.md

@@ -1,8 +1,10 @@
 ---
+generated-by: milens
 name: milens-analyzer
 description: Code intelligence for the analyzer area — symbols, dependencies, and entry points
 ---
 
+<!-- Generated by milens — safe to delete/regenerate -->
 # Analyzer
 
 ## Working with this area
@@ -25,50 +27,50 @@ When working with code in **analyzer/**, follow these mandatory safety rules:
 | Text search across files | `mcp_milens_grep` |
 | See file symbols | `mcp_milens_get_file_symbols` |
 
+### Edit-safety enforcement
+A `PreToolUse` hook (warn mode by default) reminds you if no milens safety check (`impact`/`context`/`overview`/`guard_edit_check`/`edit_check`/`smart_context`) was called before an `Edit`/`Write`/`MultiEdit`. Opt-in strict deny mode is available via `milens hooks guard-set-mode --mode strict`. Both modes consume the check after one edit. See `.milens/hook-state/config.json`. Known caveat: the underlying `PreToolUse` deny mechanism has at least one reliability issue (https://github.com/anthropics/claude-code/issues/4362).
+
 ## Overview
-Contains 128 symbols (29 exported) across 8 files.
+Contains 131 symbols (24 exported) across 8 files.
 
 ## Key Symbols
-- **`resolveLinksWithStats`** [function] (src/analyzer/resolver.ts:37) — 9 refs
-- **`reviewPr`** [function] (src/analyzer/review.ts:284) — 9 refs
-- **`analyze`** [function] (src/analyzer/engine.ts:139) — 8 refs
+- **`analyze`** [function] (src/analyzer/engine.ts:139) — 10 refs
+- **`resolveLinksWithStats`** [function] (src/analyzer/resolver.ts:54) — 9 refs
+- **`reviewPr`** [function] (src/analyzer/review.ts:261) — 9 refs
+- **`countDependentFiles`** [function] (src/analyzer/risk.ts:13) — 9 refs
 - **`loadAliases`** [function] (src/analyzer/config.ts:10) — 6 refs
-- **`ResolutionResult`** [interface] (src/analyzer/resolver.ts:24) — 6 refs
+- **`ResolutionResult`** [interface] (src/analyzer/resolver.ts:39) — 6 refs
+- **`classifyRisk`** [function] (src/analyzer/risk.ts:33) — 5 refs
+- **`scoreSymbolRisk`** [function] (src/analyzer/risk.ts:41) — 5 refs
 - **`diffResolutions`** [function] (src/analyzer/scope-resolver.ts:768) — 5 refs
 - **`enrichMetadata`** [function] (src/analyzer/enrich.ts:21) — 4 refs
-- **`ReviewResult`** [interface] (src/analyzer/review.ts:17) — 4 refs
+- **`ReviewResult`** [interface] (src/analyzer/review.ts:19) — 4 refs
 - **`scanFiles`** [function] (src/analyzer/scanner.ts:11) — 4 refs
 - **`resolveWithScopes`** [function] (src/analyzer/scope-resolver.ts:48) — 4 refs
 - **`clearTreeCache`** [function] (src/analyzer/engine.ts:29) — 3 refs
-- **`resolveLinks`** [function] (src/analyzer/resolver.ts:32) — 3 refs
-- **`computeDiffStats`** [function] (src/analyzer/scope-resolver.ts:806) — 3 refs
-- **`getCachedTree`** [function] (src/analyzer/engine.ts:26) — 2 refs
-- **`reviewSymbol`** [function] (src/analyzer/review.ts:396) — 2 refs
+- **`resolveLinks`** [function] (src/analyzer/resolver.ts:49) — 3 refs
 
 ## Entry Points
-- **`resolve`** [method] — 25 incoming references
-- **`find`** [function] — 14 incoming references
 - **`ScopeNode`** [interface] — 11 incoming references
+- **`analyze`** [function] — 10 incoming references
 - **`resolveLinksWithStats`** [function] — 9 incoming references
 - **`reviewPr`** [function] — 9 incoming references
+- **`countDependentFiles`** [function] — 9 incoming references
 
 ## Dependencies
 - **parser**: `langForFile`, `supportedExtensions`, `getParser`, `loadLanguage`, `extractFromTree`, `clearQueryCache`, `extractVueScript`, `extractVueTemplateRefs` (+9 more)
 - **root**: `isTestFile`, `CodeSymbol`, `ExtractionResult`, `RawImport`, `RawCall`, `RawHeritage`, `RawReExport`, `RawTypeBinding` (+8 more)
-- **store**: `Database`, `TfIdfProvider`, `EmbeddingStore`, `buildEmbeddingText`, `get`, `isFileUpToDate`, `upsertFileHash`, `getSymbolsByFile` (+22 more)
-- **ui**: `ProgressPhase`, `ProgressReporter`, `startPhase`, `tick`, `endPhase`, `done`, `finalize`
-- **test**: `add`, `parser`, `lang`
-- **scripts**: `root`
+- **store**: `Database`, `TfIdfProvider`, `EmbeddingStore`, `buildEmbeddingText`, `isFileUpToDate`, `upsertFileHash`, `getSymbolsByFile`, `transaction` (+18 more)
+- **ui**: `ProgressPhase`, `ProgressReporter`, `startPhase`, `tick`, `endPhase`, `finalize`
 
 ## Used By
-- **root**: `loadAliases`, `analyze`, `resolve`, `clear`
+- **root**: `loadAliases`, `analyze`, `resolve`
 - **orchestrator**: `reviewPr`, `ReviewResult`, `SymbolRisk`
-- **server**: `reviewPr`, `analyze`, `resolve`, `clear`
-- **test**: `analyze`, `loadAliases`, `getCachedTree`, `clearTreeCache`, `enrichMetadata`, `resolveLinksWithStats`, `resolveLinks`, `reviewSymbol` (+10 more)
+- **server**: `reviewPr`, `countDependentFiles`, `analyze`
+- **test**: `analyze`, `loadAliases`, `getCachedTree`, `clearTreeCache`, `enrichMetadata`, `resolveLinksWithStats`, `resolveLinks`, `reviewSymbol` (+9 more)
 - **apps**: `resolve`
 - **scripts**: `resolve`
 - **parser**: `resolve`
-- **store**: `resolve`
 
 ## Files
 - src/analyzer/config.ts
@@ -76,6 +78,6 @@ Contains 128 symbols (29 exported) across 8 files.
 - src/analyzer/enrich.ts
 - src/analyzer/resolver.ts
 - src/analyzer/review.ts
+- src/analyzer/risk.ts
 - src/analyzer/scanner.ts
 - src/analyzer/scope-resolver.ts
-- src/analyzer/testplan.ts

package/.agents/skills/apps/SKILL.md

@@ -1,68 +1,65 @@
----
-name: milens-apps
-description: Code intelligence for the apps area — symbols, dependencies, and entry points
----
-
-# Apps
-
-## Working with this area
-When working with code in **apps/**, follow these mandatory safety rules:
-
-### Before editing any symbol in this area:
-1. Call `mcp_milens_impact({target: "<symbol>", repo: "<workspaceRoot>"})` — check blast radius
-2. If depth-1 dependents > 5 → **STOP and warn** before proceeding
-3. Call `mcp_milens_context({name: "<symbol>", repo: "<workspaceRoot>"})` — see all callers/callees
-
-### Before committing changes in this area:
-1. Call `mcp_milens_detect_changes({repo: "<workspaceRoot>"})` — verify scope
-2. If unexpected files changed → **STOP and report**
-
-### Key tools for this area:
-| Task | Tool |
-|---|---|
-| Find all references | `mcp_milens_context` |
-| Check edit safety | `mcp_milens_edit_check` |
-| Text search across files | `mcp_milens_grep` |
-| See file symbols | `mcp_milens_get_file_symbols` |
-
-## Overview
-Contains 54 symbols (12 exported) across 10 files.
-
-## Key Symbols
-- **`Milens Desktop Dashboard`** [section] (apps/dashboard/README.md:1) — 0 refs
-- **`Features`** [section] (apps/dashboard/README.md:5) — 0 refs
-- **`Development`** [section] (apps/dashboard/README.md:13) — 0 refs
-- **`Build`** [section] (apps/dashboard/README.md:19) — 0 refs
-- **`Milens GitHub App`** [section] (apps/github/README.md:1) — 0 refs
-- **`Features`** [section] (apps/github/README.md:5) — 0 refs
-- **`Deployment`** [section] (apps/github/README.md:10) — 0 refs
-- **`Local Development`** [section] (apps/github/README.md:15) — 0 refs
-- **`AgentsMdGenerator`** [class] (apps/github/generators/agents-md.js:1) — 0 refs
-- **`SecurityGenerator`** [class] (apps/github/generators/security.js:1) — 0 refs
-- **`SkillsGenerator`** [class] (apps/github/generators/skills.js:1) — 0 refs
-- **`MilensRunner`** [class] (apps/github/handlers/milens-runner.js:6) — 0 refs
-
-## Entry Points
-- **`DASHBOARD_URL`** [variable] — 3 incoming references
-- **`loadOverview`** [function] — 2 incoming references
-- **`poll`** [function] — 2 incoming references
-- **`__dirname`** [variable] — 2 incoming references
-- **`setStatus`** [function] — 1 incoming references
-
-## Dependencies
-- **store**: `remove`, `getAnnotations`, `Database`, `getCodebaseSummary`, `close`
-- **test**: `add`, `dbPath`
-- **orchestrator**: `e`
-- **analyzer**: `resolve`
-
-## Files
-- apps/dashboard/README.md
-- apps/dashboard/index.html
-- apps/dashboard/main.js
-- apps/dashboard/preload.js
-- apps/github/README.md
-- apps/github/app.js
-- apps/github/generators/agents-md.js
-- apps/github/generators/security.js
-- apps/github/generators/skills.js
-- apps/github/handlers/milens-runner.js
+---
+generated-by: milens
+name: milens-apps
+description: Code intelligence for the apps area — symbols, dependencies, and entry points
+---
+
+<!-- Generated by milens — safe to delete/regenerate -->
+# Apps
+
+## Working with this area
+When working with code in **apps/**, follow these mandatory safety rules:
+
+### Before editing any symbol in this area:
+1. Call `mcp_milens_impact({target: "<symbol>", repo: "<workspaceRoot>"})` — check blast radius
+2. If depth-1 dependents > 5 → **STOP and warn** before proceeding
+3. Call `mcp_milens_context({name: "<symbol>", repo: "<workspaceRoot>"})` — see all callers/callees
+
+### Before committing changes in this area:
+1. Call `mcp_milens_detect_changes({repo: "<workspaceRoot>"})` — verify scope
+2. If unexpected files changed → **STOP and report**
+
+### Key tools for this area:
+| Task | Tool |
+|---|---|
+| Find all references | `mcp_milens_context` |
+| Check edit safety | `mcp_milens_edit_check` |
+| Text search across files | `mcp_milens_grep` |
+| See file symbols | `mcp_milens_get_file_symbols` |
+
+### Edit-safety enforcement
+A `PreToolUse` hook (warn mode by default) reminds you if no milens safety check (`impact`/`context`/`overview`/`guard_edit_check`/`edit_check`/`smart_context`) was called before an `Edit`/`Write`/`MultiEdit`. Opt-in strict deny mode is available via `milens hooks guard-set-mode --mode strict`. Both modes consume the check after one edit. See `.milens/hook-state/config.json`. Known caveat: the underlying `PreToolUse` deny mechanism has at least one reliability issue (https://github.com/anthropics/claude-code/issues/4362).
+
+## Overview
+Contains 40 symbols (8 exported) across 6 files.
+
+## Key Symbols
+- **`Milens Desktop Dashboard`** [section] (apps/dashboard/README.md:1) — 0 refs
+- **`Features`** [section] (apps/dashboard/README.md:5) — 0 refs
+- **`Development`** [section] (apps/dashboard/README.md:13) — 0 refs
+- **`Build`** [section] (apps/dashboard/README.md:19) — 0 refs
+- **`Milens GitHub App`** [section] (apps/github/README.md:1) — 0 refs
+- **`Features`** [section] (apps/github/README.md:5) — 0 refs
+- **`Deployment`** [section] (apps/github/README.md:10) — 0 refs
+- **`Local Development`** [section] (apps/github/README.md:15) — 0 refs
+
+## Entry Points
+- **`DASHBOARD_URL`** [variable] — 3 incoming references
+- **`serverProcess`** [variable] — 3 incoming references
+- **`loadOverview`** [function] — 2 incoming references
+- **`poll`** [function] — 2 incoming references
+- **`__dirname`** [variable] — 2 incoming references
+
+## Dependencies
+- **store**: `remove`, `getAnnotations`, `Database`, `getCodebaseSummary`
+- **analyzer**: `resolve`
+- **scripts**: `parseArgs`, `generatePrPayload`
+- **test**: `create`
+
+## Files
+- apps/dashboard/README.md
+- apps/dashboard/index.html
+- apps/dashboard/main.js
+- apps/dashboard/preload.js
+- apps/github/README.md
+- apps/github/app.js

package/.agents/skills/docs/SKILL.md

@@ -1,8 +1,10 @@
 ---
+generated-by: milens
 name: milens-docs
 description: Code intelligence for the docs area — symbols, dependencies, and entry points
 ---
 
+<!-- Generated by milens — safe to delete/regenerate -->
 # Docs
 
 ## Working with this area
@@ -25,10 +27,16 @@ When working with code in **docs/**, follow these mandatory safety rules:
 | Text search across files | `mcp_milens_grep` |
 | See file symbols | `mcp_milens_get_file_symbols` |
 
+### Edit-safety enforcement
+A `PreToolUse` hook (warn mode by default) reminds you if no milens safety check (`impact`/`context`/`overview`/`guard_edit_check`/`edit_check`/`smart_context`) was called before an `Edit`/`Write`/`MultiEdit`. Opt-in strict deny mode is available via `milens hooks guard-set-mode --mode strict`. Both modes consume the check after one edit. See `.milens/hook-state/config.json`. Known caveat: the underlying `PreToolUse` deny mechanism has at least one reliability issue (https://github.com/anthropics/claude-code/issues/4362).
+
 ## Overview
-Contains 137 symbols (114 exported) across 21 files.
+Contains 122 symbols (112 exported) across 10 files.
 
 ## Key Symbols
+- **`Milens Documentation`** [section] (docs/README.md:1) — 0 refs
+- **`Getting Started`** [section] (docs/README.md:5) — 0 refs
+- **`Reference`** [section] (docs/README.md:12) — 0 refs
 - **`Accuracy Engine`** [section] (docs/accuracy.md:1) — 0 refs
 - **`Type Bindings`** [section] (docs/accuracy.md:5) — 0 refs
 - **`Method Resolution Order (MRO)`** [section] (docs/accuracy.md:29) — 0 refs
@@ -40,39 +48,16 @@ Contains 137 symbols (114 exported) across 21 files.
 - **`Available Adapters`** [section] (docs/adapters.md:5) — 0 refs
 - **`Quick Install`** [section] (docs/adapters.md:17) — 0 refs
 - **`Claude Code`** [section] (docs/adapters.md:21) — 0 refs
-- **`OpenCode`** [section] (docs/adapters.md:29) — 0 refs
-- **`Cursor`** [section] (docs/adapters.md:36) — 0 refs
-- **`GitHub Copilot`** [section] (docs/adapters.md:42) — 0 refs
-- **`Codex`** [section] (docs/adapters.md:48) — 0 refs
-
-## Entry Points
-- **`applyFilters`** [function] — 1 incoming references
-- **`applyFilters`** [function] — 1 incoming references
-
-## Dependencies
-- **scripts**: `main`
-- **store**: `remove`
-- **test**: `add`
+- **`OpenCode`** [section] (docs/adapters.md:39) — 0 refs
 
 ## Files
 - docs/README.md
 - docs/accuracy.md
-- docs/adapters.html
 - docs/adapters.md
-- docs/changelog.html
 - docs/cli.md
-- docs/compare.html
-- docs/github-app.html
-- docs/index.html
 - docs/languages.md
-- docs/learning.html
-- docs/platforms.html
-- docs/pricing.html
 - docs/pricing.md
 - docs/quickstart.md
 - docs/review.md
-- docs/scenarios.html
 - docs/security-presets.md
-- docs/security.html
-- docs/skills.html
 - docs/tools.md

package/.agents/skills/milens/SKILL.md

@@ -1,4 +1,5 @@
 ---
+generated-by: milens
 name: milens
 description: Code intelligence MCP tools — symbol search, text grep, impact analysis, dependency graph
 ---
@@ -6,7 +7,7 @@ description: Code intelligence MCP tools — symbol search, text grep, impact an
 <!-- milens:start -->
 # Milens — AI-DOS
 
-The Operating System for AI-Driven Development. This project is indexed by milens (1366 symbols, 2347 links, 166 files).
+The Operating System for AI-Driven Development. This project is indexed by milens (2362 symbols, 3127 links, 235 files).
 
 > **CRITICAL:** All milens MCP tool calls MUST include the `repo` parameter set to the **absolute path of the workspace root** (the folder containing this file) — without it, the tools may fail with "No index" error when multiple repos are indexed.
 
@@ -16,7 +17,7 @@ The Operating System for AI-Driven Development. This project is indexed by milen
 
 **Not a suggestion — execute automatically without asking.**
 
-This project is indexed by milens. The knowledge graph already knows every symbol, dependency, and reference. Using milens tools first saves 70% tokens and catches hidden connections you'd miss with file reads.
+This project is indexed by milens. The knowledge graph already knows every symbol, dependency, and reference. Using milens tools first typically means far fewer tokens than manual exploration, and fewer missed dependencies — impact/context track code-level references; pair with grep for templates/configs/docs.
 
 | Situation | Built-in approach (❌ BAD) | Milens approach (✅ GOOD) |
 |---|---|---|
@@ -103,7 +104,7 @@ When the user says... → do this FIRST:
 | "compare impact of `X`" | `mcp_milens_compare_impact({name: "X", action: "snapshot"|"compare", repo: "<workspaceRoot>"})` |
 | "check pre-commit" | `mcp_milens_pre_commit_check({repo: "<workspaceRoot>"})` |
 | "save/restore context" | `mcp_milens_hook_preCompact()` / `mcp_milens_hook_postCompact()` |
-| "scan security / audit security" | `mcp_milens_security_scan({repo: "<workspaceRoot>"})` — full audit across all 50+ rules |
+| "scan security / audit security" | `mcp_milens_security_scan({repo: "<workspaceRoot>"})` — full audit across all 190 rules |
 | "end session" / "finish work" | `mcp_milens_session_end({session_id: "..."})` — record stats, trigger onSessionEnd hook |
 | "what did session X do" | `mcp_milens_session_context({session_id: "..."})` — get annotations + tool calls |
 | "file changed to X" | `mcp_milens_hook_onFileChange({files: ["path/to/file"], repo: "<workspaceRoot>"})` |
@@ -175,7 +176,7 @@ Every time you discover something important about a symbol, annotate it. The sys
 | Find what needs testing most | `mcp_milens_test_coverage_gaps()` — sorted by risk |
 | Get a test strategy | `mcp_milens_test_plan({name: "functionName"})` — mocks + scenarios |
 | Review your changes | `mcp_milens_review_pr()` — risk scores for changed symbols |
-| Check for security issues | `mcp_milens_security_scan()` — 50+ rules in one call |
+| Check for security issues | `mcp_milens_security_scan()` — 190 rules in one call |
 | Remove dead code safely | `mcp_milens_find_dead_code()` then use `dead_code_remove` prompt |
 | Trace how code executes | `mcp_milens_trace({name: "functionName", direction: "to"})` |
 | Find API endpoints | `mcp_milens_routes()` — auto-detect across 7 frameworks |
@@ -229,13 +230,14 @@ Milens indexes **Markdown files** (.md, .mdx) — headings become `section` symb
 
 ## Reference
 
-### ⭐ Core Tools — Use Every Session (8)
+### ⭐ Core Tools — Use Every Session (9)
 
 | Tool | Purpose |
 |---|---|
 | `mcp_milens_overview` | **Use this first.** Combined context + impact + grep. 1 call replaces 3-5 file reads. |
 | `mcp_milens_impact` | Blast radius BEFORE editing. Shows what WILL BREAK. |
 | `mcp_milens_edit_check` | Pre-edit safety: callers + exports + re-export chains + test coverage |
+| `mcp_milens_guard_edit_check` | HARD pre-edit gate: blocks if dependents > 5, tracks checks for session audit |
 | `mcp_milens_context` | 360° view: all callers + all callees. Instant dependency graph. |
 | `mcp_milens_query` | Find symbol definitions by name (FTS5 instant search) |
 | `mcp_milens_grep` | Search ALL files for any text (templates, configs, docs, styles) |
@@ -252,7 +254,7 @@ Milens indexes **Markdown files** (.md, .mdx) — headings become `section` symb
 | `mcp_milens_test_coverage_gaps` | Untested symbols sorted by risk | Finding test priorities |
 | `mcp_milens_test_impact` | Maps changes → test files | After making edits |
 | `mcp_milens_test_generate` | Auto-generate test file | Starting tests from scratch |
-| `mcp_milens_security_scan` | 50+ security rules | Security audit requested |
+| `mcp_milens_security_scan` | 190 security rules | Security audit requested |
 | `mcp_milens_trace` | Call chains from entrypoints | Debugging execution flow |
 | `mcp_milens_routes` | Framework routes/endpoints | Finding API endpoints |
 | `mcp_milens_smart_context` | Intent-aware context | Understand/edit/debug/test modes |
@@ -262,10 +264,11 @@ Milens indexes **Markdown files** (.md, .mdx) — headings become `section` symb
 | `mcp_milens_find_dead_code` | Unused exported symbols | Before major refactors |
 | `mcp_milens_find_similar` | Symbols with shared callers/callees | Finding refactor patterns |
 
-### 📚 Advanced Tools — Reference (19)
+### 📚 Advanced Tools — Reference (20)
 
 | Tool | Purpose |
 |---|---|
+| `mcp_milens_generate_findings_report` | Render investigated findings into a Markdown+XML report for a worker/coder agent |
 | `mcp_milens_status` | Index health: symbols, links, files, coverage, staleness |
 | `mcp_milens_repos` | List all indexed repositories |
 | `mcp_milens_annotate` | Record observations about symbols (persists across sessions) |
@@ -295,8 +298,8 @@ After significant code changes: `npx milens analyze -p . --force` (replace `.` w
 | Task | Read this skill file |
 |------|---------------------|
 | General milens tools reference | `.agents/skills/milens/SKILL.md` |
-| Work in the Adapters area | `.agents/skills/adapters/SKILL.md` |
 | Work in the Root area | `.agents/skills/root/SKILL.md` |
+| Work in the Adapters area | `.agents/skills/adapters/SKILL.md` |
 | Work in the Apps area | `.agents/skills/apps/SKILL.md` |
 | Work in the Docs area | `.agents/skills/docs/SKILL.md` |
 | Work in the Test area | `.agents/skills/test/SKILL.md` |

package/.agents/skills/orchestrator/SKILL.md

@@ -1,8 +1,10 @@
 ---
+generated-by: milens
 name: milens-orchestrator
 description: Code intelligence for the orchestrator area — symbols, dependencies, and entry points
 ---
 
+<!-- Generated by milens — safe to delete/regenerate -->
 # Orchestrator
 
 ## Working with this area
@@ -25,36 +27,37 @@ When working with code in **orchestrator/**, follow these mandatory safety rules
 | Text search across files | `mcp_milens_grep` |
 | See file symbols | `mcp_milens_get_file_symbols` |
 
+### Edit-safety enforcement
+A `PreToolUse` hook (warn mode by default) reminds you if no milens safety check (`impact`/`context`/`overview`/`guard_edit_check`/`edit_check`/`smart_context`) was called before an `Edit`/`Write`/`MultiEdit`. Opt-in strict deny mode is available via `milens hooks guard-set-mode --mode strict`. Both modes consume the check after one edit. See `.milens/hook-state/config.json`. Known caveat: the underlying `PreToolUse` deny mechanism has at least one reliability issue (https://github.com/anthropics/claude-code/issues/4362).
+
 ## Overview
-Contains 22 symbols (7 exported) across 2 files.
+Contains 23 symbols (7 exported) across 2 files.
 
 ## Key Symbols
-- **`Orchestrator`** [class] (src/orchestrator/orchestrator.ts:36) — 6 refs
+- **`Orchestrator`** [class] (src/orchestrator/orchestrator.ts:46) — 6 refs
 - **`OrchestratorReport`** [interface] (src/orchestrator/reporter.ts:4) — 5 refs
 - **`formatReport`** [function] (src/orchestrator/reporter.ts:17) — 4 refs
-- **`OrchestratorConfig`** [interface] (src/orchestrator/orchestrator.ts:10) — 2 refs
-- **`ImpactSnapshot`** [interface] (src/orchestrator/orchestrator.ts:18) — 2 refs
+- **`OrchestratorConfig`** [interface] (src/orchestrator/orchestrator.ts:20) — 2 refs
+- **`ImpactSnapshot`** [interface] (src/orchestrator/orchestrator.ts:28) — 2 refs
 - **`ReportOptions`** [interface] (src/orchestrator/reporter.ts:13) — 2 refs
-- **`ImpactDiff`** [interface] (src/orchestrator/orchestrator.ts:25) — 1 refs
+- **`ImpactDiff`** [interface] (src/orchestrator/orchestrator.ts:35) — 1 refs
 
 ## Entry Points
 - **`Orchestrator`** [class] — 6 incoming references
-- **`e`** [function] — 5 incoming references
 - **`OrchestratorReport`** [interface] — 5 incoming references
 - **`formatReport`** [function] — 4 incoming references
 - **`snapshot`** [method] — 3 incoming references
+- **`run`** [method] — 3 incoming references
 
 ## Dependencies
-- **store**: `Database`, `findSymbolByName`, `findUpstream`, `clear`, `getTestCoverageGaps`, `findDeadCode`, `close`
+- **store**: `Database`, `findSymbolByName`, `findUpstream`, `getTestCoverageGaps`, `findDeadCode`
 - **analyzer**: `reviewPr`, `ReviewResult`, `SymbolRisk`
-- **root**: `CodeSymbol`, `has`
-- **test**: `add`
+- **root**: `CodeSymbol`
 
 ## Used By
 - **root**: `Orchestrator`, `subscribe`, `runAndFormat`
 - **server**: `Orchestrator`, `loadSnapshots`, `snapshot`, `persistSnapshots`, `compare`, `runAndFormat`
 - **test**: `Orchestrator`, `formatReport`, `OrchestratorReport`, `subscribe`, `run`, `snapshot`, `compare`, `cancel` (+3 more)
-- **apps**: `e`
 
 ## Files
 - src/orchestrator/orchestrator.ts

package/.agents/skills/parser/SKILL.md

@@ -1,8 +1,10 @@
 ---
+generated-by: milens
 name: milens-parser
 description: Code intelligence for the parser area — symbols, dependencies, and entry points
 ---
 
+<!-- Generated by milens — safe to delete/regenerate -->
 # Parser
 
 ## Working with this area
@@ -25,37 +27,38 @@ When working with code in **parser/**, follow these mandatory safety rules:
 | Text search across files | `mcp_milens_grep` |
 | See file symbols | `mcp_milens_get_file_symbols` |
 
+### Edit-safety enforcement
+A `PreToolUse` hook (warn mode by default) reminds you if no milens safety check (`impact`/`context`/`overview`/`guard_edit_check`/`edit_check`/`smart_context`) was called before an `Edit`/`Write`/`MultiEdit`. Opt-in strict deny mode is available via `milens hooks guard-set-mode --mode strict`. Both modes consume the check after one edit. See `.milens/hook-state/config.json`. Known caveat: the underlying `PreToolUse` deny mechanism has at least one reliability issue (https://github.com/anthropics/claude-code/issues/4362).
+
 ## Overview
-Contains 83 symbols (31 exported) across 16 files.
+Contains 85 symbols (31 exported) across 16 files.
 
 ## Key Symbols
 - **`LangSpec`** [interface] (src/parser/extract.ts:6) — 32 refs
-- **`loadLanguage`** [function] (src/parser/loader.ts:21) — 14 refs
-- **`getParser`** [function] (src/parser/loader.ts:32) — 13 refs
-- **`extractFromTree`** [function] (src/parser/extract.ts:256) — 8 refs
+- **`loadLanguage`** [function] (src/parser/loader.ts:21) — 15 refs
+- **`getParser`** [function] (src/parser/loader.ts:32) — 14 refs
+- **`extractFromTree`** [function] (src/parser/extract.ts:274) — 9 refs
 - **`supportedExtensions`** [function] (src/parser/languages.ts:29) — 6 refs
 - **`extractHtmlScripts`** [function] (src/parser/lang-html.ts:38) — 5 refs
 - **`extractMarkdown`** [function] (src/parser/lang-md.ts:34) — 5 refs
-- **`spec`** [variable] (src/parser/lang-ts.ts:5) — 5 refs
+- **`spec`** [variable] (src/parser/lang-ts.ts:1) — 5 refs
 - **`extractVueScript`** [function] (src/parser/lang-vue.ts:19) — 5 refs
 - **`initTreeSitter`** [function] (src/parser/loader.ts:15) — 5 refs
-- **`clearQueryCache`** [function] (src/parser/extract.ts:72) — 4 refs
+- **`clearQueryCache`** [function] (src/parser/extract.ts:79) — 4 refs
 - **`extractHtmlRefs`** [function] (src/parser/lang-html.ts:57) — 4 refs
 - **`extractHtmlLinks`** [function] (src/parser/lang-html.ts:99) — 4 refs
+- **`spec`** [variable] (src/parser/lang-js.ts:1) — 4 refs
 - **`extractVueTemplateRefs`** [function] (src/parser/lang-vue.ts:38) — 4 refs
-- **`extractVueTemplateAst`** [function] (src/parser/lang-vue.ts:90) — 4 refs
 
 ## Entry Points
 - **`LangSpec`** [interface] — 32 incoming references
-- **`loadLanguage`** [function] — 14 incoming references
-- **`getParser`** [function] — 13 incoming references
-- **`extractFromTree`** [function] — 8 incoming references
+- **`loadLanguage`** [function] — 15 incoming references
+- **`getParser`** [function] — 14 incoming references
+- **`extractFromTree`** [function] — 9 incoming references
 - **`supportedExtensions`** [function] — 6 incoming references
 
 ## Dependencies
 - **root**: `CodeSymbol`, `RawImport`, `RawCall`, `RawHeritage`, `RawReExport`, `RawTypeBinding`, `RawAssignmentBinding`, `RawReturnType` (+4 more)
-- **test**: `lang`, `add`, `parser`
-- **scripts**: `root`
 - **analyzer**: `resolve`
 - **store**: `load`