v1.2.0: fixes, Claude Code drift, simpler kit

.claude/agents/_agent-template.md → .claude/agents/_agent-template.md.template

@@ -2,7 +2,7 @@
 name: agent-name
 description: One or two sentences. State WHEN to use this agent, not just what it is. The main session routes by matching a request against this text, so name the triggers, domains, and verbs that should land here. Note what this agent is NOT for, to keep routing clean.
 tools: Read, Glob, Grep
-model: sonnet
+model: inherit
 ---
 
 # Agent Name

.claude/agents/analyst.md

@@ -2,7 +2,7 @@
 name: analyst
 description: Metrics, KPIs, and data-backed analysis. Use for pipeline and revenue figures, dashboards, forecasts, trend and cohort analysis, variance investigation, and any question that needs numbers decomposed and compared rather than guessed.
 tools: Read, Write, Edit, Glob, Grep, Bash, WebFetch, WebSearch
-model: sonnet
+model: inherit
 ---
 
 # Analyst

.claude/agents/comms-meetings.md

@@ -2,7 +2,7 @@
 name: comms-meetings
 description: Email triage, meeting preparation, daily and weekly briefings, communication drafting, and follow-up tracking. Use for "what's in my inbox", "prep me for my meeting with X", "draft a reply to Y", "what am I waiting on", and "who owes me a response".
 tools: Read, Write, Edit, Glob, Grep, Bash, WebSearch
-model: sonnet
+model: inherit
 ---
 
 # Comms & Meetings

.claude/agents/crm-relationships.md

@@ -2,7 +2,7 @@
 name: crm-relationships
 description: Contacts, pipeline, and relationship tracking. Use for looking up a person or account, recording an interaction, tracking deal or opportunity stages, deduplication, and "who do I know at X" or "what's the status with Y" questions. Works from vault files by default; uses a CRM integration if one is enabled.
 tools: Read, Write, Edit, Glob, Grep, Bash
-model: sonnet
+model: inherit
 ---
 
 # CRM & Relationships

.claude/agents/daily-copilot.md

@@ -2,7 +2,7 @@
 name: daily-copilot
 description: The everyday work partner. Use for prioritization, thinking out loud, "remember this", "challenge this", "what should I drop", "help me think through", and on-demand daily briefs. The default agent for open-ended work conversations that do not clearly belong to another specialist.
 tools: Read, Write, Edit, Glob, Grep, Bash, WebSearch
-model: sonnet
+model: inherit
 ---
 
 # Daily Copilot

.claude/agents/first-principles.md

@@ -2,7 +2,7 @@
 name: first-principles
 description: Rigorous abstract reasoning. Use to reason a problem from scratch, decompose a concept to its fundamentals, calibrate a prediction, run scenario analysis under deep uncertainty, or synthesize opposing views. For thinking about which framework even applies, not applying a known one to a decision.
 tools: Read, Glob, Grep, WebFetch, WebSearch
-model: sonnet
+model: inherit
 ---
 
 # First Principles

.claude/agents/principal-engineer.md

@@ -2,7 +2,7 @@
 name: principal-engineer
 description: Senior engineering voice for technical decisions. Use to design a system, review code or architecture, choose between technologies, assess technical debt, or do a security review. Reasons against modern best practices. Advice only; produces recommendations, not commits.
 tools: Read, Glob, Grep, Bash, WebFetch, WebSearch
-model: sonnet
+model: inherit
 ---
 
 # Principal Engineer

.claude/agents/second-brain.md

@@ -2,7 +2,7 @@
 name: second-brain
 description: Knowledge base management. Use for taking notes, creating and retrieving meeting records, research synthesis, finding what was discussed with a person or on a topic, and keeping the vault healthy (wikilinks, frontmatter, archival). Read-only by default; writes happen on explicit intent.
 tools: Read, Write, Edit, Glob, Grep, Bash
-model: sonnet
+model: inherit
 ---
 
 # Second Brain

.claude/agents/strategy-advisor.md

@@ -2,7 +2,7 @@
 name: strategy-advisor
 description: Strategic sparring partner. Use when bringing a plan, idea, or decision for critique, or for pre-mortems, steelmans, second-order analysis, and "should I do X". Challenges assumptions and surfaces what is being missed. Advice only; does not execute, write files, or send anything.
 tools: Read, Glob, Grep, WebFetch, WebSearch
-model: sonnet
+model: inherit
 ---
 
 # Strategy Advisor

.claude/settings.json

@@ -16,7 +16,7 @@
   "hooks": {
     "SessionStart": [
       {
-        "matcher": "startup|resume|compact",
+        "matcher": "startup|resume|clear|compact|fork",
         "hooks": [
           { "type": "command", "command": "bash \"$CLAUDE_PROJECT_DIR/hooks/hook-session-start.sh\"" }
         ]

.claude/skills/add-agent/SKILL.md

@@ -13,10 +13,10 @@ Create a new agent so the assistant grows with the user's work. This is the main
    - **Name** — kebab-case, e.g. `recruiting` or `product-research`.
    - **Purpose and triggers** — what it owns, and the phrases or domains that should route to it. This becomes the `description`, which is what actually drives routing.
    - **Tools** — the minimum it needs (4 to 6). Advice-only agents get read tools only: `Read, Glob, Grep, WebFetch, WebSearch`.
-   - **Model** — `sonnet` is the default; `opus` for genuinely hard reasoning.
+   - **Model**: `inherit` is the default (the agent follows the session's model); pin `sonnet` for a cheap, high-volume agent or `opus` for one that always needs hard reasoning.
 2. **Check for overlap.** Read the existing agents in `.claude/agents/`. If the new agent overlaps an existing one, say so and suggest either editing the existing agent or sharpening both descriptions so routing stays clean.
-3. **Scaffold.** Copy `.claude/agents/_agent-template.md` to `.claude/agents/<name>.md` and fill it in: frontmatter, Purpose, When to use / when not to, Output format, Safety.
-4. **Register it.** Add a one-line row for the new agent to `.claude/agents/_orchestrator-reference.md` and to the routing section of `CONSTITUTION.md`.
+3. **Scaffold.** Copy `.claude/agents/_agent-template.md.template` to `.claude/agents/<name>.md` and fill it in: frontmatter, Purpose, When to use / when not to, Output format, Safety.
+4. **Register it.** Add a one-line routing rule for the new agent to the Agent Architecture & Routing section of `CONSTITUTION.md`. (`_orchestrator-reference.md` describes coordination patterns; it is not a list of agents.)
 5. **Test routing.** Suggest two or three prompts that should land on the new agent, and one that should *not* (to confirm it does not over-trigger).
 
 ## Principles

.claude/skills/mcp-debug/SKILL.md

@@ -101,7 +101,7 @@ Status: <from health check>
 
 1. Run the status check (Mode 1) for the full server list.
 
-2. If tool-usage telemetry exists (some projects log to `data/performance/session-log.jsonl` or similar), summarise the last 7 days:
+2. If you keep tool-usage telemetry (this kit does not ship any), summarise the last 7 days:
    - Tools called per server.
    - Error rate per server.
    - Most-used and least-used servers.

.claude/skills/memory-debt/SKILL.md

@@ -82,7 +82,7 @@ Final summary: X applied, Y skipped, Z marked obsolete, W silent fixes.
 
 - **Never auto-apply.** Always present each item for approval before modifying any file.
 - **Read before writing.** If the content already exists, mark APPLIED silently — never duplicate.
-- **MEMORY.md line cap.** If the constitution sets a limit (commonly ~200 lines), and writing the proposal would breach it, suggest a topic file in `memory/topics/` and a one-line index pointer instead of inlining.
+- **MEMORY.md size cap.** MEMORY.md loads every session, so keep it under about 12KB. If writing the proposal would push it past that, suggest a topic file in `memory/topics/` and a one-line index pointer instead of inlining.
 - **Diary edits are minimal.** Only flip checkboxes and update Memory Debt table status. Never rewrite proposal text.
 - **Oldest first.** Process chronologically so the longest-overdue debt clears first.
 

.github/SECURITY.md

@@ -4,7 +4,7 @@ Compabob runs entirely on your own machine, under your own Claude Code subscript
 
 ## Supported version
 
-Only the latest `main` is supported. The kit is a rolling template, not a versioned product. If you found an issue, check it still reproduces on the current `main` before reporting.
+Releases are tagged (see the CHANGELOG), but only the latest `main` is supported: fixes land there, not on old tags. If you found an issue, check it still reproduces on the current `main` before reporting.
 
 ## Reporting a vulnerability
 

CHANGELOG.md

@@ -6,8 +6,36 @@ All notable changes to Compabob are recorded here. Format follows
 
 ## [Unreleased]
 
+## [1.2.0] - 2026-09-25
+
+Maintenance release: the modules and skills that shipped since 1.1.1, fixes
+for bugs found in a full audit, the Claude Code changes since June, and a
+smaller kit. Safe to pull with `./update.sh`: your `vault/`, `memory/`,
+`config/user.config.yaml`, `.mcp.json`, and `.env` are untouched, and any new
+config flags are added for you.
+
+**If you edited kit files**, `update.sh` may ask you to merge the ones this
+release also changed: every agent in `.claude/agents/`, `.claude/settings.json`,
+`CLAUDE.md`, the `add-agent`, `mcp-debug`, and `memory-debt` skills, `setup.sh`,
+`update.sh`, everything in `scripts/` and `docs/`, and the proactive, telegram,
+transcribe, dictation, and integrations modules. `git diff --stat v1.1.1 v1.2.0`
+lists them all. Personas and the other skills are unchanged.
+
 ### Added
 
+- **`dictation` module**: a cleanup endpoint your dictation app (e.g. Handy)
+  posts to. It fixes punctuation and fillers, and rewrites words it keeps
+  getting wrong using a glossary that grows itself from recurring corrections.
+  OpenAI-compatible upstream (Groq by default, Ollama for fully local).
+- **`/weekly-review` skill**: end-of-week review of what got done, what
+  slipped, decisions, commitments, and next week's top three (#16).
+- **`researcher` persona preset** for academia and R&D (#19).
+- **Spanish README intro** at `docs/i18n/README.es.md` (#23).
+- **Linux and WSL troubleshooting** section in the README (#14).
+- **`scripts/migrate-config.sh`**: adds config keys a kit update introduces,
+  inside the right block, with a `.bak`; never changes a value you set.
+  `update.sh` runs it, and `scripts/init.sh` warns when keys are missing.
+- **`scripts/lib/common.sh`**: shared output helpers and `set_config_flag`.
 - **`transcribe` module (highly encouraged)** — one-hotkey local call recording +
   transcription. A toggle script drives `ffmpeg` to capture audio and
   `mlx-whisper` to transcribe it on-device: first run records, second run stops,
@@ -52,24 +80,60 @@ All notable changes to Compabob are recorded here. Format follows
   as PENDING / STALE / OBSOLETE / APPLIED, then in `apply` mode walks through
   approvals one by one. Closes the loop between `/reflect` (proposes) and
   `memory/`.
-- **Tool-scope-guard hook** (`hooks/hook-tool-scope-guard.sh` +
-  `tool_scope_check.py` + `tool_scopes.yaml`) — a context-scoped PreToolUse
-  guard. Reasoning stays unconstrained; only the outward action surface
-  (sends, mutations, external writes) is bounded per execution context.
-  Pre-wired contexts: `interactive` (allow-all), `pulse` (overnight headless,
-  no sends), `morning-briefing` (read inboxes, write vault), `email-triage`
-  (label only, no send), `news-digest` (fetch + write only). Context is
-  selected via `$CLAUDE_CONTEXT` env var or `/tmp/claude-context-<session>`
-  file. Fail-open by design; blocks are logged to
-  `data/performance/tool-scope-blocks.jsonl`. Opt-in: add the hook entry to
-  `.claude/settings.json` under `PreToolUse` to activate.
-
-### Documented
-
-- **Choosing the Claude model** — new README section explains how to swap
-  the default model per session (`/model` or `--model`), per project
-  (`.claude/settings.json` → `"model"`), or globally (`ANTHROPIC_MODEL` env
-  var), with a pointer to the Anthropic model list. Closes #9.
+
+### Changed
+
+- **Agents follow your model.** Every agent uses `model: inherit` instead of
+  `sonnet`, so switching models in Claude Code switches the agents too. On a
+  Pro plan, see the README for pinning them back to Sonnet. Scheduled modules
+  still run on Sonnet.
+- **Install Claude Code with the native installer**
+  (`curl -fsSL https://claude.ai/install.sh | bash`); npm stays documented as a
+  fallback (Node 22+). Docs links point at code.claude.com.
+- **README model section rewritten** (closes #9): Claude Code's default is now
+  Opus 5.5 on every paid plan; how to change it per session, per project, or
+  globally.
+- **Integrations**: pins bumped (`@playwright/mcp` 0.0.82, `scrapling-fetch-mcp`
+  0.2.4, `mcp-server-time` 2026.8.18, `exa-mcp-server` 3.4.1, which needs
+  Node 20+). Gmail and Calendar now point at Claude's own connectors.
+- **SessionStart hook** also runs after `/clear` and in forked sessions.
+- **Config template**: flags for `crm_merge` and `lead_pipeline`; persona list
+  shows all six.
+- **`transcribe` and `dictation`** install their Python dependency into a
+  module venv (Homebrew Python refuses a global `pip install`); `record.py`
+  switches to its venv on its own, so hotkeys keep working.
+- **`update.sh`** says what to do on a local-only branch or detached HEAD
+  instead of blaming the network.
+- **`scripts/init.sh`** checks for Python 3.10+ and counts only real agents.
+
+### Removed
+
+- **Tool-scope guard hook** (never released): it was not wired into
+  `settings.json`, needed `jq`, and matched no context the kit sets.
+- **Stub modules** `extra-agents`, `team`, and `whatsapp`, which were README
+  files only. Their reasoning now lives in the "Not built" section of
+  `modules/README.md`; the roadmap is issue #12.
+- **`mcp-server-fetch`** from the integrations catalog (Claude Code's built-in
+  WebFetch covers it) and the empty `google` category.
+- **Dead config keys**: `preferences.timezone`, `extra_agents`, `team`. Existing
+  configs keep them; nothing reads them.
+
+### Fixed
+
+- **Weekly review ran on Saturday on macOS**: launchd counts Sunday as 0, so
+  `Weekday 6` was Saturday. Now Friday.
+- **Scheduled runs could not find `claude`** under launchd's minimal PATH
+  (`proactive`, `telegram`).
+- **The agent template registered as a real agent**. It is now
+  `.claude/agents/_agent-template.md.template`.
+- **`install-integrations.sh` left an empty `.mcp.json`** on runs that added
+  nothing.
+- **Docs claimed `config/` is git-ignored**; only `config/user.config.yaml` is.
+- **`/add-agent` and the customization guide** pointed at
+  `_orchestrator-reference.md` as an agent registry; routing lives in
+  `CONSTITUTION.md`.
+- **Telegram README** said the hook "blocks until you approve"; the hook always
+  blocks, and you send with `send.sh` yourself.
 
 ## [1.1.1] — 2026-05-24
 
@@ -202,6 +266,8 @@ Initial public release at [github.com/chacosoldier/compabob](https://github.com/
 - `docs/architecture.md`, `docs/onboarding.md`, `docs/customization-guide.md`,
   `docs/how-to-improve-memory.md`.
 
-[Unreleased]: https://github.com/chacosoldier/compabob/compare/v1.1.0...HEAD
+[Unreleased]: https://github.com/chacosoldier/compabob/compare/v1.2.0...HEAD
+[1.2.0]: https://github.com/chacosoldier/compabob/compare/v1.1.1...v1.2.0
+[1.1.1]: https://github.com/chacosoldier/compabob/compare/v1.1.0...v1.1.1
 [1.1.0]: https://github.com/chacosoldier/compabob/compare/v1.0.0...v1.1.0
 [1.0.0]: https://github.com/chacosoldier/compabob/releases/tag/v1.0.0

CLAUDE.md

@@ -10,4 +10,4 @@ At the start of a session, read `memory/MEMORY.md`. It records the user's name,
 
 ## Working directory
 
-The knowledge base is `vault/`. Persistent facts are in `memory/`. Your own configuration is in `.claude/` and `config/`. The directories `vault/`, `memory/`, and `config/` are the user's own data and live outside the kit's version control. Everything else is the kit, and updates upstream.
+The knowledge base is `vault/`. Persistent facts are in `memory/`. Your own configuration is in `.claude/` and `config/`. `vault/`, `memory/`, and `config/user.config.yaml` are the user's own data and live outside the kit's version control. Everything else is the kit, and updates upstream.

CONTRIBUTING.md

@@ -6,7 +6,7 @@ This kit is built to be changed. "Contributing" here means two things: making it
 
 The kit assumes you will edit it. Nothing here is sacred.
 
-- **Agents**: edit any file in `.claude/agents/`. Add a new one with the `/add-agent` skill, or copy `.claude/agents/_agent-template.md`.
+- **Agents**: edit any file in `.claude/agents/`. Add a new one with the `/add-agent` skill, or copy `.claude/agents/_agent-template.md.template`.
 - **Skills**: each directory in `.claude/skills/` is one slash command. Copy an existing one to make a new workflow.
 - **Hooks**: `hooks/` holds safety guards and lifecycle scripts. They are wired in `.claude/settings.json`. Add, remove, or disable any of them.
 - **Constitution**: `CONSTITUTION.md` is the behavior contract. Change it to match how you actually work.