feat: close all audit gaps — tests, spec docs, config, verification, website polish

- INTERNALS.md: hooks table expanded from 5 to 8 entries (added session-start, teammate-idle, task-completed)
- Skills: renamed all 15 SKILL.md → index.md + updated 19 files with references
- Commands: added /maxsim:execute-phase alias, Ctrl+G tips in 11 Plan Mode commands
- Tests: 6 new test files (check-update, notification-sound, stop-sound, statusline, version, install-hooks fixture)
- Workflows: per-task verification in execute.md, per-task TSV logging, verify-phase.md deliverable checks
- Config: permissions block in settings.json, ensureDiscussionCategory in discussions.ts, worktree config options
- Website: renamed planning-directory.md → github-project-board.md, cleaned .planning/ prohibition text
- Package: added test:e2e script, lint fixes (biome compliance)
- Commands: debug.md auto-trigger note, progress.md overdue detection, init.md brownfield support

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: maystudios

CHANGELOG.md

@@ -16,7 +16,7 @@
 
 ### Features
 
-* add executable Tier 2 Agent Teams patterns to execute.md and maxsim-batch SKILL.md ([8460d8a](https://github.com/maystudios/maxsimcli/commit/8460d8a6d4e95522b6e90a9e10cda6a34ae14320))
+* add executable Tier 2 Agent Teams patterns to execute.md and maxsim-batch index.md ([8460d8a](https://github.com/maystudios/maxsimcli/commit/8460d8a6d4e95522b6e90a9e10cda6a34ae14320))
 * add Tier selection logic to debug.md and plan-research.md workflows ([ef9590e](https://github.com/maystudios/maxsimcli/commit/ef9590e8ee9aa976607e875c5601e5c80a48ef2f))
 * **cli:** add resolve-max-agents command with task complexity scaling ([cd16494](https://github.com/maystudios/maxsimcli/commit/cd16494e3d23678ec3c4ae2e51d43fcfc0945a4d))
 

INTERNALS.md

@@ -263,7 +263,7 @@ Exported functions:
 
 ## Hook System
 
-Five hook scripts compiled by tsdown into `dist/assets/hooks/*.cjs` and
+**Seven** hook scripts (plus statusLine) compiled by tsdown into `dist/assets/hooks/*.cjs` and
 installed to `.claude/maxsim/hooks/` at `npx maxsimcli` time.
 
 | Hook script                    | Event       | Purpose                                      |
@@ -273,6 +273,9 @@ installed to `.claude/maxsim/hooks/` at `npx maxsimcli` time.
 | `maxsim-notification-sound.cjs`| Notification| Play sound when Claude asks a question       |
 | `maxsim-stop-sound.cjs`        | Stop        | Play sound when Claude finishes              |
 | `maxsim-capture-learnings.cjs` | Stop        | Append session commits to MEMORY.md          |
+| `maxsim-session-start.cjs`     | SessionStart | Inject MEMORY.md + TSV + git log context at session start |
+| `maxsim-teammate-idle.cjs`     | TeammateIdle | Check pending tasks before teammate goes idle (exit 2 blocks) |
+| `maxsim-task-completed.cjs`    | TaskCompleted| Run test/build/lint gates before task completion (exit 2 blocks) |
 
 All hooks share `hooks/shared.ts`:
 - `CLAUDE_DIR = '.claude'` constant.
@@ -290,12 +293,17 @@ All hooks share `hooks/shared.ts`:
 ```json
 {
   "hooks": {
-    "SessionStart": [{ "hooks": [{ "type": "command", "command": "node \"...maxsim-check-update.cjs\"" }] }],
+    "SessionStart": [
+      { "hooks": [{ "type": "command", "command": "node \"...maxsim-check-update.cjs\"" }] },
+      { "hooks": [{ "type": "command", "command": "node \"...maxsim-session-start.cjs\"" }] }
+    ],
     "Notification":  [{ "hooks": [{ "type": "command", "command": "node \"...maxsim-notification-sound.cjs\"" }] }],
     "Stop": [
       { "hooks": [{ "type": "command", "command": "node \"...maxsim-stop-sound.cjs\"" }] },
       { "hooks": [{ "type": "command", "command": "node \"...maxsim-capture-learnings.cjs\"" }] }
-    ]
+    ],
+    "TeammateIdle": [{ "hooks": [{ "type": "command", "command": "node \"...maxsim-teammate-idle.cjs\"" }] }],
+    "TaskCompleted": [{ "hooks": [{ "type": "command", "command": "node \"...maxsim-task-completed.cjs\"" }] }]
   },
   "statusLine": { "command": "node \"...maxsim-statusline.cjs\"" },
   "env": { "CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1" }
@@ -403,7 +411,7 @@ directly — it is copied into `dist/assets/templates/` during the build.
 
 ### Skills (14)
 
-Located at `templates/skills/`, one subdirectory per skill with a `SKILL.md`.
+Located at `templates/skills/`, one subdirectory per skill with an `index.md`.
 All follow Anthropic's skill convention: YAML frontmatter with `name` and
 `description`, body under 500 lines, no `@` imports.
 
@@ -529,6 +537,8 @@ User: /maxsim:execute 3
 
 5. User approves (ExitPlanMode)
 
+> **Tip:** Users can press **Ctrl+G** while reviewing a plan to edit it in their default text editor before approving.
+
 6. Workflow: spawn executor agents (Agent tool, isolation: worktree)
    → one worktree per task: .maxsim-worktrees/phase-3-task-{id}/
    → branch per task: maxsim/phase-3-task-{id}

README.md

@@ -350,7 +350,7 @@ Project config lives in `.claude/maxsim/config.json`, created during `/maxsim:in
 
 ## Skills
 
-Skills are reusable prompt modules that agents load on demand. Each skill is a `SKILL.md` file with YAML frontmatter and a markdown body.
+Skills are reusable prompt modules that agents load on demand. Each skill is an `index.md` file with YAML frontmatter and a markdown body.
 
 ### Built-in Skills (14)
 

docs/spec/claude-md-guide.md

@@ -622,7 +622,7 @@ When MaxsimCLI spawns a subagent (Executor, Planner, Researcher, Verifier), the
 Skills do not explicitly import CLAUDE.md — it is already in context when a skill runs. A skill can reference instructions from CLAUDE.md without repeating them:
 
 ```markdown
-# SKILL.md for deployment skill
+# index.md for deployment skill
 
 Deploy following the project commit conventions (see CLAUDE.md) and the
 verification protocol (see .claude/rules/verification-protocol.md).

docs/spec/parallel-execution-guide.md

@@ -1031,8 +1031,8 @@ Do workers need to share findings mid-task?
 
 **Local MaxsimCLI Reference:**
 - `/c/Development/cli/maxsim/docs/claude-own-skills-ref/batch.md` — Batch parallel orchestration skill
-- `/c/Development/cli/maxsim/docs/superpowers-reference/skills/dispatching-parallel-agents/SKILL.md` — Parallel agent dispatch patterns
-- `/c/Development/cli/maxsim/docs/superpowers-reference/skills/subagent-driven-development/SKILL.md` — Sequential subagent development with review gates
+- `/c/Development/cli/maxsim/docs/superpowers-reference/skills/dispatching-parallel-agents/index.md` — Parallel agent dispatch patterns
+- `/c/Development/cli/maxsim/docs/superpowers-reference/skills/subagent-driven-development/index.md` — Sequential subagent development with review gates
 
 **Community:**
 - [Claude Code Sub-Agents: Parallel vs Sequential Patterns](https://claudefa.st/blog/guide/agents/sub-agent-best-practices)

docs/spec/plan-mode-guide.md

@@ -1086,5 +1086,5 @@ claude --permission-mode plan -p "Create a migration plan for the auth system"
 - [plannotator — Visual Plan Review for Claude Code](https://plannotator.ai/)
 - [plannotator GitHub Repository](https://github.com/backnotprop/plannotator)
 - [Short Notes — Claude Code Settings: Default to Plan Mode and Plans Directory](https://notes.ghinda.com/post/claude-code-settings-default-to-plan-mode-and-plans-directory)
-- [MAXSIM writing-plans skill](/c/Development/cli/maxsim/docs/superpowers-reference/skills/writing-plans/SKILL.md)
-- [MAXSIM executing-plans skill](/c/Development/cli/maxsim/docs/superpowers-reference/skills/executing-plans/SKILL.md)
+- [MAXSIM writing-plans skill](/c/Development/cli/maxsim/docs/superpowers-reference/skills/writing-plans/index.md)
+- [MAXSIM executing-plans skill](/c/Development/cli/maxsim/docs/superpowers-reference/skills/executing-plans/index.md)

docs/spec/self-improvement-guide.md

@@ -584,7 +584,7 @@ The metric is the most critical decision in setting up an autoresearch loop. A b
 
 | What to Improve | Metric Command | Direction |
 |-----------------|---------------|---------|
-| Skill token usage | `wc -w < .claude/skills/target-skill/SKILL.md` | Lower |
+| Skill token usage | `wc -w < .claude/skills/target-skill/index.md` | Lower |
 | Command response time | Custom timing harness | Lower |
 | Verification pass rate | Count of pass/fail over N runs | Higher |
 | Agent retry count | Parse stop hook logs for retry events | Lower |
@@ -737,7 +737,7 @@ The autoresearch pattern was designed for code optimization. MaxsimCLI uses it a
 ### What Changes in Project Management Context
 
 **The target is process, not code.** Instead of modifying implementation files, the loop modifies:
-- Skill instructions (`.claude/skills/*/SKILL.md`)
+- Skill instructions (`.claude/skills/*/index.md`)
 - Command definitions (`.claude/commands/maxsim/*.md`)
 - Agent prompts (`.claude/agents/*.md`)
 - Workflow definitions (`.claude/workflows/*.md`)
@@ -772,7 +772,7 @@ Goal: Reduce the number of agent retries required during phase execution
 Verify:  node maxsim-tools.cjs stats --retries --last-phase
 Guard:   npx vitest run
 Direction: lower_is_better
-Scope:   .claude/skills/verification/SKILL.md
+Scope:   .claude/skills/verification/index.md
          .claude/agents/executor.md
 ```
 
@@ -994,7 +994,7 @@ Mode:       Bounded (30 iterations)
 **Setup:**
 ```
 Goal:       Reduce executor agent retry rate for verification tasks
-Scope:      .claude/skills/verification/SKILL.md
+Scope:      .claude/skills/verification/index.md
             .claude/agents/executor.md
 Verify:     node maxsim-tools.cjs stats --retries --last-20-sessions
 Guard:      npx vitest run
@@ -1035,7 +1035,7 @@ Mode:       Bounded (20 iterations)
 ```
 Goal:       Reduce the average number of tasks added/removed after planning
 Scope:      .claude/commands/maxsim/plan.md
-            .claude/skills/roadmap-writing/SKILL.md
+            .claude/skills/roadmap-writing/index.md
 Verify:     gh api repos/owner/repo/issues --jq '[.[] | select(.labels[].name == "task") | .number] | length'
 Guard:      (none — this metric has no binary pass/fail guard)
 Direction:  lower_is_better (fewer post-plan task changes = better upfront planning)

docs/spec/self-improvement-research.md

@@ -286,16 +286,16 @@ Replace the current `[NEEDS RESEARCH]` section with the architecture described i
 | Source | URL/Path |
 |--------|----------|
 | autoresearch repo | `github.com/uditgoenka/autoresearch` (cloned to `/tmp/autoresearch/`) |
-| autoresearch SKILL.md | `claude-plugin/skills/autoresearch/SKILL.md` |
+| autoresearch index.md | `claude-plugin/skills/autoresearch/index.md` |
 | autoresearch loop protocol | `claude-plugin/skills/autoresearch/references/autonomous-loop-protocol.md` |
 | autoresearch results logging | `claude-plugin/skills/autoresearch/references/results-logging.md` |
 | autoresearch core principles | `claude-plugin/skills/autoresearch/references/core-principles.md` |
 | Superpowers reference | `docs/superpowers-reference/` |
 | Superpowers research | `docs/superpowers-research.md` |
 | MaxsimCLI self-improvement guide | `docs/spec/self-improvement-guide.md` |
 | MaxsimCLI memory system guide | `docs/spec/memory-system-guide.md` |
-| MaxsimCLI verification skill | `templates/skills/verification/SKILL.md` |
-| MaxsimCLI project-memory skill | `templates/skills/project-memory/SKILL.md` |
+| MaxsimCLI verification skill | `templates/skills/verification/index.md` |
+| MaxsimCLI project-memory skill | `templates/skills/project-memory/index.md` |
 | MaxsimCLI capture-learnings hook | `packages/cli/src/hooks/maxsim-capture-learnings.ts` |
 | MaxsimCLI self-improvement ref | `templates/references/self-improvement.md` |
 | MaxsimCLI verification patterns | `templates/references/verification-patterns.md` |

docs/spec/skills-specification.md

@@ -932,7 +932,7 @@ Not preloaded. User-invocable on-demand. Activates when any of the 4 autonomous
 
 ### Estimated Line Count
 
-~169 lines (SKILL.md body, excluding reference files)
+~169 lines (index.md body, excluding reference files)
 
 ---
 
@@ -1010,20 +1010,20 @@ No circular references exist in the target state.
 New skills require new directories under `templates/skills/`:
 
 ```
-templates/skills/verification/SKILL.md       (new — merge)
-templates/skills/github-operations/SKILL.md  (new — merge)
-templates/skills/research/SKILL.md           (new — merge)
-templates/skills/project-memory/SKILL.md     (new)
+templates/skills/verification/index.md       (new — merge)
+templates/skills/github-operations/index.md  (new — merge)
+templates/skills/research/index.md           (new — merge)
+templates/skills/project-memory/index.md     (new)
 ```
 
 ### Files to Update
 
 ```
-templates/skills/tdd/SKILL.md                       (update See also)
-templates/skills/systematic-debugging/SKILL.md      (fix "5-Step" → "6-Step", update See also)
-templates/skills/roadmap-writing/SKILL.md           (remove MAXSIM Integration section)
-templates/skills/code-review/SKILL.md               (add SPEC COMPLIANCE dimension)
-templates/skills/using-maxsim/SKILL.md              (update skills table)
+templates/skills/tdd/index.md                       (update See also)
+templates/skills/systematic-debugging/index.md      (fix "5-Step" → "6-Step", update See also)
+templates/skills/roadmap-writing/index.md           (remove MAXSIM Integration section)
+templates/skills/code-review/index.md               (add SPEC COMPLIANCE dimension)
+templates/skills/using-maxsim/index.md              (update skills table)
 templates/agents/AGENTS.md                          (update preloaded skills references)
 templates/agents/executor.md                        (update: evidence-collection → verification)
 templates/agents/researcher.md                      (update: research-methodology → research)