wade-cli 0.35.3__tar.gz → 0.48.1__tar.gz

wade_cli-0.48.1/.gitattributes

@@ -0,0 +1,4 @@
+# wade:knowledge:start
+KNOWLEDGE.md merge=union
+KNOWLEDGE.ratings.jsonl merge=union
+# wade:knowledge:end

{wade_cli-0.35.3 → wade_cli-0.48.1}/.github/workflows/ci.yml

@@ -60,6 +60,8 @@ jobs:
         run: ./scripts/test.sh tests/unit/
       - name: Run integration tests
         run: ./scripts/test.sh tests/integration/
+      - name: Run concurrency lane (shared git/remote state)
+        run: ./scripts/test-concurrency.sh
       - name: Run top-level CLI contract smoke
         run: ./scripts/test.sh tests/test_cli_basics.py
 

{wade_cli-0.35.3 → wade_cli-0.48.1}/.gitignore

@@ -28,11 +28,13 @@ htmlcov/
 # AI tool directories (wade-managed — never committed in this repo)
 .agents/
 .claude/
+.codex/
 .copilot/
 .cursor/
 .gemini/
 .github/hooks/
 .github/skills
+.wade/
 CLAUDE.md
 PLAN.md
 PR-SUMMARY.md

{wade_cli-0.35.3 → wade_cli-0.48.1}/.wade.yml

@@ -12,23 +12,28 @@ ai:
   plan:
     effort: xhigh
     yolo: true
+  implement:
+    yolo: true
   deps:
     tool: claude
     model: claude-sonnet-5
     mode: headless
     effort: high
+    timeout: 900
   review_plan:
     tool: claude
     model: claude-sonnet-5
     mode: headless
     effort: high
     enabled: true
+    timeout: 900
   review_implementation:
     tool: claude
     model: claude-sonnet-5
     mode: headless
     effort: high
     enabled: true
+    timeout: 900
   review_batch:
     tool: claude
     model: claude-sonnet-5
@@ -61,5 +66,3 @@ hooks:
   copy_to_worktree:
   - .env
   - .wade.yml
-  - KNOWLEDGE.md
-  - KNOWLEDGE.ratings.yml

{wade_cli-0.35.3 → wade_cli-0.48.1}/AGENTS.md

@@ -16,8 +16,8 @@ Two distinct worlds interact in this codebase. Always be clear which one you are
 | **the WADE repo** / **this project** | This source repository — `src/wade/`, `templates/`, `tests/`, `scripts/` |
 | **inited project** / **target project** | Any third-party repo that has run `wade init` to adopt the workflow |
 | **skill templates** | Markdown files in `templates/skills/` — the source of truth, part of the WADE repo |
-| **installed skills** | Copies (or symlinks) of skill templates placed in a project's `.claude/skills/` by `wade init` |
-| **AGENTS.md pointer** | A short `## Git Workflow` block that `wade init` injects into an inited project's `AGENTS.md` |
+| **installed skills** | Copies (or symlinks) of skill templates placed in a project's `.claude/skills/` **per session by worktree bootstrap** (`bootstrap_worktree`), not by `wade init` |
+| **AGENTS.md pointer** | A short `## Git Workflow` block that **worktree bootstrap** injects into an inited project's `AGENTS.md` per session, not by `wade init` |
 
 **This `AGENTS.md` governs development of WADE itself.** Skills, the pointer, and the progressive disclosure architecture are all *outputs* of WADE — artifacts installed into inited projects, not rules for developing WADE.
 
@@ -53,15 +53,34 @@ uv run python scripts/auto_version.py patch # Version bump (patch/minor/major)
 
 ```
 CLI Layer      ->  can import: services, models, config, logging, ui
-Service Layer  ->  can import: providers, crossby (AI tool adapters), git, db, models, config, logging
+Service Layer  ->  can import: providers, crossby (AI tool adapters), git, models, config, logging
 Provider Layer ->  can import: models, config, logging  (NO service imports)
 Git Layer      ->  can import: models, config, logging  (NO service imports)
-DB Layer       ->  can import: models, logging           (NO config imports)
 Models Layer   ->  can import: nothing (leaf dependency)
 ```
 
 No circular dependencies. Models are pure data. Services orchestrate. **Never import a higher layer from a lower layer.**
 
+> **The `db/` package is unused scaffolding** — intentionally left out of the
+> layering rules above. No code path under `services/` or `cli/` writes
+> session/worktree/PR rows, so its sole reader
+> (`implementation_service/cleanup._preserve_session_data`, a single
+> `SessionRepository.get_by_worktree_path` call) always gets an empty result and
+> falls back to directory-presence detection. Real persisted state lives in
+> GitHub (PR/issue body markers, labels) and worktree files, **not** SQLite. The
+> code still exists but is inert; removal is tracked as #357 C5.
+
+> **Deterministic git-hook install/reconcile/build logic lives in `git/hooks.py`**
+> (git layer), not `skills/installer.py`. Packaged template-asset loaders
+> (prompt/skill/hook templates) live in `utils/templates.py`, a leaf module.
+> Leaf `utils/` modules import nothing from wade, so any layer — including the
+> git layer — may import them without breaking the rules above.
+> `skills` itself is a lower utility/template layer — services may import it for
+> **skill-file management** (`install_skills`, the `*_SKILLS` registries,
+> `ensure_knowledge_merge_attributes`, gitignore/cross-tool constants), never the
+> reverse. Those residual `service -> skills` imports are sanctioned, not
+> layering violations.
+
 AI tool adapters are not part of this repo — they live in the external [`crossby`](https://github.com/ivanviragine/crossby) package (`pyproject.toml`). See `docs/dev/architecture.md` for what moved there.
 
 CLI modules are thin dispatch — they parse flags via Typer, then call service methods. Business logic lives in `services/`, not in `cli/`.
@@ -100,6 +119,8 @@ Everything in this repo exists in one of two worlds:
 
 When developing WADE, **only touch the left column**. Always edit `templates/skills/<name>/SKILL.md` directly — never edit files inside `.claude/skills/` (those are symlinks in this repo, copies in inited projects).
 
+The right-column artifacts (installed skills, the `## Git Workflow` pointer, and the Claude/Cursor allowlists) are produced **per session by worktree bootstrap** (`bootstrap_worktree` in `implementation_service/bootstrap.py`, invoked by `wade implement`/`wade plan`/`wade review` and by a standalone `wade task deps`), *not* by `wade init`. `wade init` writes only `.wade.yml`, optional provider/knowledge files, and the `.wade/` manifest.
+
 > Skills system deep dive (symlinks, pointer markers, installation lifecycle): see `docs/dev/skills-system.md`
 
 ## Conventions
@@ -130,6 +151,13 @@ Before considering any work complete:
 - [ ] **`templates/skills/`** — updated if agent-facing rules changed (plan-session for planning, implementation-session for implementation, review-pr-comments-session for reviews)
 - [ ] **Commit** — uses conventional-commit prefix
 
+Note: for inited projects, the doc update pass is now an explicit, mandatory
+closing step in the installed `implementation-session` and
+`review-pr-comments-session` skills (`{doc_update_step}` /
+`templates/skills/_partials/doc-update-step.md`) — not something an agent has
+to remember on its own. This repo's own checklist above is unaffected; it
+still governs what WADE-repo changes require here.
+
 > Full 10-item checklist, documentation rules, feedback loop, and correction-driven docs: see `docs/dev/documentation-policies.md`
 
 ## Detailed Reference

{wade_cli-0.35.3 → wade_cli-0.48.1}/CHANGELOG.md

@@ -4,6 +4,265 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Conventional Commits](https://conventionalcommits.org/).
 
+## [v0.48.1] — 2026-08-17
+
+### Documentation
+
+- tighten session-skill communication — terse reporting and unambiguous exit/complete signaling (#436) (0384f72)
+
+## [v0.48.0] — 2026-08-17
+
+### Features
+
+- add config-driven bot-review trigger command with optional auto-trigger (#432) (fc2ff0d)
+
+## [v0.47.0] — 2026-08-17
+
+### Features
+
+- adopt crossby 0.24.0 refreshed AI-tool model catalogs (#434) (3842332)
+
+## [v0.46.1] — 2026-08-16
+
+### Bug Fixes
+
+- propagate Codex worktree launch context through sessions (#429) (032daa2)
+
+## [v0.45.10] — 2026-08-16
+
+### Bug Fixes
+
+- resolve branch by issue number so retitled issues resume their PR/plan (#428) (c840b5e)
+
+## [v0.45.9] — 2026-08-16
+
+### Bug Fixes
+
+- apply resolved autonomy mode to resumed implementation sessions (#427) (30cc8c9)
+
+## [v0.45.8] — 2026-08-16
+
+### Chores
+
+- default implementation sessions to yolo, matching plan/review_batch (#424) (43054af)
+
+## [v0.45.7] — 2026-08-16
+
+### Bug Fixes
+
+- preserve user-set ai.implement keys on re-init (#425) (71a25db)
+
+## [v0.45.6] — 2026-08-16
+
+### Bug Fixes
+
+- forward resolved permission mode to child sessions when explicit (#426) (fbdd9c9)
+
+## [v0.45.5] — 2026-08-16
+
+### Bug Fixes
+
+- drop done.require_review: false bypass hint from done review-refusal output (#422) (5424a00)
+
+## [v0.45.4] — 2026-08-15
+
+### Bug Fixes
+
+- _pull_main_after_merge untracked-collision backup is never restored on retry failure (#420) (de59eb7)
+
+## [v0.45.3] — 2026-08-14
+
+### Bug Fixes
+
+- resolve worktree/branch by issue number + project-scope UX (#417) (6158a41)
+
+## [v0.45.2] — 2026-08-13
+
+### Refactoring
+
+- shared typed worktree model for list_worktrees() consumers (#416) (69931c7)
+
+## [v0.45.1] — 2026-08-13
+
+### Bug Fixes
+
+- merge_pr hard-fails on recoverable "Head branch is out of date" (stale PR head sync) (#415) (d81a658)
+
+## [v0.45.0] — 2026-08-13
+
+### Features
+
+- support a configurable base branch for plans and implementation (#377) (7bbde95)
+
+## [v0.44.3] — 2026-08-12
+
+### Refactoring
+
+- move deterministic git-hook logic out of skills.installer (service-layer boundary) (#412) (d93468d)
+
+## [v0.44.2] — 2026-08-12
+
+### Bug Fixes
+
+- allow writes to system temp dirs through worktree/plan write-guards (#410) (81fcf74)
+
+## [v0.44.1] — 2026-08-12
+
+### Bug Fixes
+
+- wade implement startup catchup silently aborts on knowledge-store-migration untracked_conflict, leaving the session on a stale base (#408) (9968583)
+
+## [v0.44.0] — 2026-08-12
+
+### Features
+
+- make WADE session output report-by-exception (terse on success) (#402) (9f96c94)
+
+## [v0.43.0] — 2026-08-11
+
+### Features
+
+- allow AI-tool memory writes through worktree/plan write-guards (#388) (338a31c)
+
+## [v0.42.6] — 2026-08-11
+
+### Bug Fixes
+
+- always surface the resolved permission mode and honor it in review sessions (#380) (01beb8a)
+
+## [v0.42.5] — 2026-08-11
+
+### Bug Fixes
+
+- a timed-out headless review loses everything and is likely to time out again (#406) (1050ff9)
+
+## [v0.42.4] — 2026-08-10
+
+### Bug Fixes
+
+- restore nested-session guard for Antigravity CLI in wade implement (#405) (8d16c06)
+
+## [v0.42.3] — 2026-08-10
+
+### Bug Fixes
+
+- don't declare review complete when bot hasn't reviewed the latest commit (#404) (f36ed80)
+
+## [v0.42.2] — 2026-08-10
+
+### Bug Fixes
+
+- `done` cannot tell a clean review from one that never ran (#400) (8788d52)
+
+## [v0.42.1] — 2026-08-10
+
+### Chores
+
+- remove dead pre-push install wrappers superseded by the batch git-hook API (#399) (e39517a)
+
+## [v0.42.0] — 2026-08-09
+
+### Features
+
+- enforce conventional-commit PR & issue titles across creation and done (#393) (c3f13d5)
+
+## [v0.41.1] — 2026-08-09
+
+### Bug Fixes
+
+- allow /dev/ device writes in plan-mode write guard (#398) (adc6c61)
+
+### Other Changes
+
+- Lock-retry for `git stash push` is git-version-dependent (follow-up to #357 C3) (#396) (27e0eb9)
+
+## [v0.41.0] — 2026-08-08
+
+### Features
+
+- re-inject session context on start/resume/compaction (#351) (#391) (8a9c0ed)
+
+### Other Changes
+
+- wade review crashes with Rich MarkupError when reviewer feedback contains bracket markup (#395) (2a224a6)
+
+## [v0.40.0] — 2026-08-08
+
+### Features
+
+- honor a dedicated ai.review_pr_comments config for the auto-launched review session (#390) (d6b920e)
+
+### Other Changes
+
+- H3: Knowledge lifecycle — worktree-local, reviewed and merged through the PR (#386) (26af39f)
+
+## [v0.39.3] — 2026-08-07
+
+### Bug Fixes
+
+- bound implementation-session review loop with a code-enforced 2-pass cap (#385) (3ed4f04)
+
+### Other Changes
+
+- E4: Repo-quality gates — pre-commit + commit-msg git hooks, and PostToolUse feedback (#381) (123727a)
+- E2: Plan-phase enforcement (auto-validate + plan Stop guard) (#378) (92e6b43)
+- E1: Make `done` the authoritative completion gate (+ pre-push backstop, + realign Stop) (#375) (2445c55)
+
+## [v0.39.2] — 2026-08-05
+
+### Chores
+
+- stop tracking KNOWLEDGE.md/KNOWLEDGE.ratings.yml in git (c6a78b1)
+
+### Other Changes
+
+- H4: Drift & layering — AGENTS.md, architecture.md, crossby hook comments, raw subprocess in services (#365) (3ae0ff3)
+
+## [v0.39.1] — 2026-08-05
+
+### Bug Fixes
+
+- allow shell reads outside the worktree, block only writes (#373) (d9dc842)
+
+## [v0.39.0] — 2026-08-05
+
+### Features
+
+- offer only tool-valid reasoning-effort levels in wade init (#371) (cb6458a)
+
+## [v0.38.1] — 2026-08-04
+
+### Bug Fixes
+
+- accept permission_mode key in config validation (#369) (cdeb48e)
+
+### Other Changes
+
+- H2: Git & remote-state hygiene — never lose work, never act on a stale read (+ retire `merge_strategy: direct`) (#362) (6e55893)
+- H1: Hook-layer correctness — capability matrix + fail-closed containment (crossby 0.13) (#363) (623805f)
+
+## [v0.38.0] — 2026-08-03
+
+### Features
+
+- add a documentation update pass to the implementation and pr-comments sessions (#361) (96b6b6d)
+
+### Other Changes
+
+- E: Context budget — slim the skills, say each rule once, load reference just-in-time (#364) (670cc29)
+
+## [v0.37.0] — 2026-08-01
+
+### Features
+
+- modernize AI-tool hook integration via crossby runtime contract (#320) (76e8ca2)
+
+## [v0.36.0] — 2026-07-25
+
+### Features
+
+- announce headless review timeout so orchestrators don't kill it early (#347) (1af5754)
+
 ## [v0.35.3] — 2026-07-25
 
 ### Chores