docs: record the process skills, the /sm re-scope, and the ignore rules

claude/CLAUDE.md gains a Process skills section so the six cross-cutting skills
are reached for deliberately rather than discovered by accident, and states the
rule they exist to enforce: a ticket is Done only when its negative criteria are
met and each names the symbol enforcing it.

CHANGELOG records the additions and the /sm re-scope. README's skill count
updated for the new skills.

Author: olehsvyrydov

CHANGELOG.md

@@ -4,6 +4,21 @@ This file records all notable changes to the project. Versioning roughly follows
 
 ## [Unreleased]
 
+### Added
+- **Five cross-cutting process skills** — roles cover *who* does the work; these cover *how* it is decided, checked and paid for. `fid-lifecycle` (backlog → design doc/epic → tickets → Done, without orphans or two disagreeing records), `verify-landed` (prove a change exists and does something — a green build is not evidence, because removing nothing breaks nothing), `review-tier` (how much multi-agent review to buy and how to scope it), `model-selection` (which model tier to spend — the test is whether a cheap verifier exists downstream), `research-method` (investigate so the result is trustworthy; publish so it survives a hostile reader). Listed in `claude/commands/agents.md` and `claude/CLAUDE.md`.
+- **`answer-audit`** — adversarially audit a retrieval-grounded answer, assuming it is wrong until each claim is proven verbatim against a source passage. Targets the failure class a grounding judge structurally cannot see: an answer minus its qualifier is still *entailed* by its sources, so "is this grounded?" passes it.
+- **`/ext` — Browser Extension Developer**, with references for MV3 architecture, injection and extraction, permissions and security, tabs and groups, long-running work, surfaces, and build/test.
+- **Theia references** for `/arch`, `/ui` and `/fe`, and a **Rust reference** for `/be`.
+- **`references/board-integrity.md`** for `/sm` — six integrity checks with checkable answers, plus retrospection by continuous capture.
+
+### Changed
+- **`/sm` re-scoped from ceremony to board integrity.** The role read "you are the conductor of the orchestra — every agent reports to you", which is unfalsifiable, so nothing ever contradicted it and in practice it was never invoked. It is now accountable for orphan work items, plan-vs-tracker divergence, backlog items whose work has moved on, Done-without-negative-criteria, parked-without-a-trigger, and epics claiming completion over open children. Output is counts and named items, never a verdict.
+- **Retrospectives are continuous, not scheduled.** Capture one question at merge; distil the running page per epic. A fortnightly retro reconstructs the past from memory and yields platitudes — the durable lessons are found mid-work, while the mechanism is still in your hands.
+- **`workflow-engine`** gains the tracking-location rule: a local folder or an external tracker is a convenience call by project size, and never both for the same thing.
+
+### Fixed
+- `.gitignore` now covers machine-local agent state (`.praxis/`, `.aidevteam/`) and archived internal discussions (`docs/archive/`). This repository is public, and those transcripts carry product strategy and private-project detail.
+
 ## [5.1.0] — 2026-06-14
 
 A consolidation and hygiene release that finishes the framework repositioning: the on-demand specialist roster is now backed by vendor-neutral reference libraries, two dormant agents are wired to commands, the team roster and knowledge-backend naming are reconciled across the docs, and the dashboard product, the RAG subsystem, and the personal migration scripts are removed.

README.md

@@ -2,7 +2,7 @@
 
 ![Claude Code](https://img.shields.io/badge/Claude%20Code-ready-6E56CF) ![Cursor](https://img.shields.io/badge/Cursor-ready-111111) ![Kiro](https://img.shields.io/badge/Kiro-ready-9B59B6) ![VS Code](https://img.shields.io/badge/VS%20Code-ready-007ACC) ![License: MIT](https://img.shields.io/badge/License-MIT-green) ![Stars](https://img.shields.io/github/stars/olehsvyrydov/AI-development-team?style=social)
 
-> **A reusable AI agent-team framework for your editor.** ~29 specialist agent skills (a 15-agent core team + optional specialists) plus a `workflow-engine` and an **enforced, proportional workflow** with approval gates — *process, not prompts*. Open-source, no lock-in; works in **Claude Code, Cursor, Kiro & VS Code**, free by default.
+> **A reusable AI agent-team framework for your editor.** ~35 specialist agent skills (a 15-agent core team + optional specialists) plus a `workflow-engine` and an **enforced, proportional workflow** with approval gates — *process, not prompts*. Open-source, no lock-in; works in **Claude Code, Cursor, Kiro & VS Code**, free by default.
 
 ```
 /po → /arch → /secops → [/fin] → [/legal] → [/ui] → /fe | /be → /rev → /qa + /e2e → /verify

claude/CLAUDE.md

@@ -15,9 +15,22 @@ A team of ~29 specialist AI agents (a 15-agent core team + optional specialists)
 - **Gates** (set as labels in the ledger): `ARCH_APPROVED`, `SECOPS_APPROVED`, `DESIGN_APPROVED`, `APPROVAL_GATE`, `CODE_REVIEWED`, `PERF_OK`, `VERIFIED`, `RELIABILITY_OK`. Security gates are a safety override — never skipped for being "small."
 - Full detail: `TEAM_WORKFLOW.md` (on-demand). Definition: `workflow/workflow.yaml`.
 
+## Process skills — load these, they are not optional extras
+Beside the role agents, six **cross-cutting process skills** cover how work is decided, checked and paid for. They self-trigger from their descriptions; know they exist so you reach for them deliberately.
+- **`workflow-engine`** — the gate contract. Consult first, on every task.
+- **`fid-lifecycle`** — moving work Backlog → investigation → design doc / epic → tickets → Done without orphans or two disagreeing records. Delete the backlog item once its design doc exists; one epic per design doc, titled to correspond; every ticket parented to it and listed in its registry; **the tracker is authoritative for status** when the two disagree.
+- **`verify-landed`** — after any behaviour-changing edit, and before claiming a finding fixed or a criterion met. A green build is not evidence your change exists: *removing nothing breaks nothing*, so a no-op edit compiles and passes every test.
+- **`review-tier`** — how much code review to buy and how to scope it. Free gates green first; on a follow-up round review *the fix and the prior findings*, not the whole PR again.
+- **`model-selection`** — which model tier to spend. The test is whether a cheap verifier exists downstream: if a mistake would be caught mechanically, go cheaper; if this output is the last line of defence, spend.
+- **`research-method`** — for questions that need measuring rather than deciding, and for writing results up so they survive a hostile reader.
+
+**The rule they enforce**, because it is the most expensive lesson so far: a ticket is Done only when its *negative* criteria are met — the guard, the refusal, the "cannot bypass" — and each names the **symbol** that enforces it. An audit of one project's 30 open tickets found 13 with an unmet criterion, nearly always the negative case, several treated as complete because the feature existed. Happy paths ship; the things that stop bad outcomes do not, unless someone checks.
+
 ## Tickets & docs — no Jira required
 Default is **file-based**: markdown tickets (**Backlog.md**) + a markdown knowledge base (Obsidian-compatible). Jira/Confluence and other backends (a knowledge-graph backend) are **optional MCP overlays**, enabled only in `workflow.yaml`.
 
+Where progress is tracked — a local folder or an external tracker — is a **convenience call**, decided by project size. Never run both for the same thing: two trackers become two answers to "what is the state?", and the stale one is always the one someone reads.
+
 ## Memory (optional)
 Native files by default. Optional OSS memory MCP (OpenMemory / mem0). No memory backend is required.