code-atlas-mcp 0.2.0.dev1__tar.gz → 0.4.0__tar.gz

{code_atlas_mcp-0.2.0.dev1 → code_atlas_mcp-0.4.0}/.atlasignore

@@ -17,3 +17,7 @@ migrations/
 testdata/
 fixtures/
 *.sql
+
+# Generated knowledge-vault landing page (atlas dream) — derived, re-rendered
+# every run; indexing it would just be indexing yesterday's dream report.
+docs/HOME.md

code_atlas_mcp-0.4.0/.claude/commands/dream-mode.md

@@ -0,0 +1,97 @@
+---
+description:
+  Dream mode (code-atlas edition) — consolidate drafts across the repo inbox, the harness memory dir, and the global
+  vault, using atlas dream's deterministic report.
+---
+
+# Dream Mode — Code Atlas Edition
+
+Project-local override of the global `dream-mode` command (per that command's own "extend, don't duplicate" convention).
+Everything in the global command still applies — philosophy, disposition table, KEEP/MERGE/PROMOTE/DROP semantics,
+output format. This file only changes **where the inventory comes from** and adds the knowledge-vault-specific promotion
+venues and archive convention that code-atlas's unified graph makes possible.
+
+## Why this repo needs its own version
+
+In code-atlas, repo-inbox drafts (`docs/inbox/`), the harness memory dir, and (if configured) the global vault are **all
+Notes in the same Memgraph graph** — one `atlas dream` (or the `knowledge_health` MCP tool) call produces a single
+deterministic report spanning all three, instead of three separate manual audits. Capture-time routing is best-effort
+(write wherever's cheapest mid-session); this command is the authoritative consolidation pass across all of them, same
+as the global command's philosophy — just with a graph-backed inventory instead of a directory listing.
+
+## 1. Inventory (replaces the global command's step 1)
+
+- Run `atlas dream --json` (or call the `knowledge_health` MCP tool) instead of manually listing memory-dir files. It
+  returns, across every configured vault (this repo's `docs/`, the harness memory dir, and any
+  `[knowledge] extra_vaults`):
+  - `inbox_count` / `inbox_paths` — draft notes awaiting disposition
+  - `orphan_notes` — notes with no `[[wikilinks]]` in or out (disconnected from the note graph)
+  - `dangling_links` — `[[wikilink]]`/`derived_from`/`supersedes` references whose target doesn't exist (typo'd slug, or
+    the target was deleted/renamed)
+  - `duplicate_ids` — two files in the same vault resolved to the same note uid (only one survives in the graph; the
+    report is the only way to see the collision)
+  - `similar_pairs` — high-embedding-similarity note pairs (candidates for MERGE)
+  - `promotion_candidates` — similar pairs that span _different_ projects (repo-specific finding that recurs elsewhere →
+    candidate for the global vault)
+  - `memory_index_issues` — MEMORY.md entries with no matching file, or files not indexed
+- Still read `MEMORY.md` directly for the human-facing index (byte size, entry count) — the report's
+  `memory_index_issues` only flags _consistency_, not size.
+- List promotion venues exactly as the global command does (skills, rules, CLAUDE.md, inline comments, docs/) **plus**:
+  - **`docs/notes/` / `docs/decisions/`** — durable zettels in this repo's own vault (`kind: note` / `kind: decision`
+    frontmatter, per `docs/SCHEMA.md`)
+  - **The global vault** — cross-project generalizations (`promotion_candidates` above is the mechanical signal for this
+    venue)
+- Do not modify anything yet.
+
+## 2. Classify
+
+Same disposition table as the global command (KEEP / MERGE / PROMOTE / CLAUDE.md / inline comment / docs / DROP), with
+two additions specific to the graph-unified vault:
+
+| Disposition                | When                                                                                                                       | Action at execute step                                                                                                                                                         |
+| -------------------------- | -------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| **PROMOTE → global vault** | A `promotion_candidates` hit, or a finding with no repo-specific anchor that clearly generalizes.                          | `git mv` the file to the global vault's path; add `promoted_from: <old-uid>` to frontmatter so provenance survives the move.                                                   |
+| **DROP → archive stub**    | Content worth a provenance pointer but not worth keeping in full (superseded decision, stale draft with historical value). | See §4's archive-stub convention below — this is a refinement of the global command's plain DROP, available because the graph can query `DERIVED_FROM`/`SUPERSEDES` afterward. |
+
+Use `duplicate_ids` and `dangling_links` as blocking findings — resolve them before proceeding to promotions, since a
+duplicate-id file will silently overwrite the other on next index regardless of what you decide for either one
+individually.
+
+## 3. Plan and confirm
+
+Same as the global command — present the disposition table, wait for approval. Extend the table's venue column with
+`global vault` / `archive stub` where applicable.
+
+## 4. Execute
+
+Same ordering as the global command (promotions → merges → deletions → refresh index → backlinks), plus:
+
+- **Archive-stub convention** (instead of outright deletion for anything with lasting reference value): reduce the file
+  to a 2-3 line summary, set `archived: true` in frontmatter, and `git mv` it to an `archive/` subdirectory of its
+  vault. Point the stub at the commit hash where the full content last existed (`git log --oneline -1 -- <path>` before
+  the edit). The stub stays indexed (still searchable, still a real Note) — only the full content leaves the live vault.
+  Never do this for a note still referenced by live `DERIVED_FROM`/`SUPERSEDES` edges without first redirecting them.
+- **Provenance on MERGE/PROMOTE**: write `derived_from: [<source-uid>, ...]` (and `supersedes: [<uid>]` when the new
+  note fully replaces an old one) in the survivor's frontmatter — these become real `DERIVED_FROM`/`SUPERSEDES` graph
+  edges on next index, so provenance is queryable (`MATCH (n:Note)-[:DERIVED_FROM]->(m) ...`), not just prose.
+- **After execution, re-run `atlas dream`** (not just the manual checks the global command lists) to confirm
+  `duplicate_ids`/`dangling_links` actually cleared and the promoted/ merged notes appear where expected — the report is
+  cheap and catches mistakes the manual grep-based verification in the global command's step 5 can miss (e.g. a
+  `derived_from` slug typo that silently fails to resolve).
+
+## 5. Verify
+
+Everything in the global command's step 5, plus:
+
+- `atlas dream --json` again — `duplicate_ids` and `dangling_links` should be empty (or only contain items explicitly
+  deferred, noted as such in the punch list).
+- For each `PROMOTE → global vault` row: confirm the note is indexed under the global vault's project name
+  (`atlas status` or `hybrid_search(mode="knowledge")`).
+
+## Output format
+
+Same punch list as the global command, plus one line:
+
+```
+Vault findings resolved: <duplicate_ids before>→<after>, <dangling_links before>→<after>
+```

code_atlas_mcp-0.4.0/.claude/rules/knowledge.md

@@ -0,0 +1,30 @@
+# Knowledge Vault Pointer
+
+This repo's `docs/` directory is a knowledge vault indexed into the code-atlas graph alongside the code — a `Note` node
+per frontmattered file, linked to code entities and to each other. Full conventions: `docs/SCHEMA.md`.
+
+This file is a stable pointer, not a status digest — it doesn't change as notes are added; look the state up when you
+need it instead of reading it here.
+
+## Finding things
+
+- `hybrid_search(query, mode="knowledge")` — ranks notes/docs above code, for why/decision/gotcha-shaped questions.
+  `mode="blended"` (default) still includes them, ranked slightly below code. `mode="code"` excludes them.
+- `get_context(uid)` on a code entity returns its linked notes/docs via `DOCUMENTS` edges.
+- Cypher: `Note` nodes have `uid = "{project}:note:{slug}"`, `kind` in `draft|note|decision`, `tags`, `docstring` (full
+  body). `LINKS_TO` is the wikilink graph; `DERIVED_FROM`/`SUPERSEDES` are dream-mode provenance.
+
+## Writing things
+
+- Quick findings mid-session: drop a frontmattered file in `docs/inbox/` (see SCHEMA.md for the frontmatter shape) — or
+  the Claude Code memory dir if it's a private/machine-local observation, not something the repo should carry. Both are
+  equivalent draft piles; don't agonize over which.
+- A `remember` MCP write tool for this is planned (Phase 2) but not built yet.
+- Durable notes in `docs/notes/`/`docs/decisions/` are produced by dream-mode consolidation only — don't write there
+  directly.
+
+## Status
+
+Phases 0–1 of the knowledge-convergence architecture are implemented (vault parsing, schema, search integration).
+Capture tooling (`remember`), anchor staleness, and dream-mode consolidation are not built yet — see
+`.tasks/research/2026-07-11-knowledge-convergence-architecture.md` for the full roadmap.

code_atlas_mcp-0.4.0/.gitattributes

@@ -0,0 +1,84 @@
+# Default
+# ==================
+*                   text=auto eol=lf
+
+# Python Source files
+# =================
+*.pxd               text diff=python
+*.py                text diff=python
+*.py3               text diff=python
+*.pyw               text diff=python
+*.pyx               text diff=python
+*.pyz               text diff=python
+*.pyi               text diff=python
+
+# Python Binary files
+# =================
+*.db                binary
+*.p                 binary
+*.pkl               binary
+*.pickle            binary
+*.pyc               binary export-ignore
+*.pyo               binary export-ignore
+*.pyd               binary
+
+# Jupyter notebook
+# =================
+*.ipynb             text
+
+# ML models
+# =================
+*.h5                filter=lfs diff=lfs merge=lfs -text
+*.onnx              filter=lfs diff=lfs merge=lfs -text
+*.model             filter=lfs diff=lfs merge=lfs -text
+*.msgpack           filter=lfs diff=lfs merge=lfs -text
+*.pb                filter=lfs diff=lfs merge=lfs -text
+*.pt                filter=lfs diff=lfs merge=lfs -text
+*.pth               filter=lfs diff=lfs merge=lfs -text
+pytorch_model.bin   filter=lfs diff=lfs merge=lfs -text
+tokenizer.json      filter=lfs diff=lfs merge=lfs -text
+unigram.json        filter=lfs diff=lfs merge=lfs -text
+
+# Data files
+# =================
+*.csv               filter=lfs diff=lfs merge=lfs -text
+*.tsv               filter=lfs diff=lfs merge=lfs -text
+*.parquet           filter=lfs diff=lfs merge=lfs
+
+# Presentation files
+# =================
+*.pptx              filter=lfs diff=lfs merge=lfs -text
+*.word              filter=lfs diff=lfs merge=lfs -text
+*.xlsx              filter=lfs diff=lfs merge=lfs -text
+*.xls               filter=lfs diff=lfs merge=lfs -text
+*.pdf               filter=lfs diff=lfs merge=lfs -text
+
+# Archives
+# =================
+*.7z                filter=lfs diff=lfs merge=lfs -text
+*.br                filter=lfs diff=lfs merge=lfs -text
+*.gz                filter=lfs diff=lfs merge=lfs -text
+*.tar               filter=lfs diff=lfs merge=lfs -text
+*.tgz               filter=lfs diff=lfs merge=lfs -text
+*.tar.gz            filter=lfs diff=lfs merge=lfs -text
+*.zip               filter=lfs diff=lfs merge=lfs -text
+*.rar               filter=lfs diff=lfs merge=lfs -text
+
+# Image files
+# =================
+*.jpg           filter=lfs diff=lfs merge=lfs -text
+*.jpeg          filter=lfs diff=lfs merge=lfs -text
+*.png           filter=lfs diff=lfs merge=lfs -text
+*.gif           filter=lfs diff=lfs merge=lfs -text
+*.webp          filter=lfs diff=lfs merge=lfs -text
+*.bmp           filter=lfs diff=lfs merge=lfs -text
+*.svg           filter=lfs diff=lfs merge=lfs -text
+*.tiff          filter=lfs diff=lfs merge=lfs -text
+
+# Other
+# =================
+# Windows - keep CRLF
+*.exe           filter=lfs diff=lfs merge=lfs -text
+*.bat            text eol=crlf
+*.cmd            text eol=crlf
+*.ps1            text eol=crlf

{code_atlas_mcp-0.2.0.dev1 → code_atlas_mcp-0.4.0}/.github/workflows/release.yml

@@ -31,19 +31,22 @@ jobs:
       contents: write
 
     steps:
+      - uses: actions/create-github-app-token@v1
+        id: app-token
+        with:
+          app-id: ${{ vars.RELEASE_APP_ID }}
+          private-key: ${{ secrets.RELEASE_APP_PRIVATE_KEY }}
+
       - uses: actions/checkout@v6
         with:
           fetch-depth: 0
-          ref: ${{ github.ref_name }}
-
-      - name: Pin to triggering SHA
-        run: git reset --hard ${{ github.sha }}
+          token: ${{ steps.app-token.outputs.token }}
 
       - name: Semantic Release
         id: release
         uses: python-semantic-release/python-semantic-release@v10
         with:
-          github_token: ${{ secrets.GITHUB_TOKEN }}
+          github_token: ${{ steps.app-token.outputs.token }}
           git_committer_name: "github-actions[bot]"
           git_committer_email: "github-actions[bot]@users.noreply.github.com"
           force: ${{ inputs.force || '' }}
@@ -53,7 +56,7 @@ jobs:
         if: steps.release.outputs.released == 'true'
         uses: python-semantic-release/publish-action@v10
         with:
-          github_token: ${{ secrets.GITHUB_TOKEN }}
+          github_token: ${{ steps.app-token.outputs.token }}
           tag: ${{ steps.release.outputs.tag }}
 
     outputs:
@@ -67,7 +70,7 @@ jobs:
   # ------------------------------------------------------------------- #
   publish-pypi:
     needs: release
-    if: needs.release.outputs.released == 'true'
+    if: needs.release.outputs.released == 'true' && needs.release.outputs.is_prerelease != 'true'
     runs-on: ubuntu-latest
     environment:
       name: pypi
@@ -89,6 +92,15 @@ jobs:
         with:
           python-version: "3.14-dev"
 
+      - name: Verify version matches tag
+        run: |
+          TAG="${{ needs.release.outputs.tag }}"
+          TOML_VERSION=$(python -c "import tomllib; print('v' + tomllib.load(open('pyproject.toml','rb'))['project']['version'])")
+          if [ "$TAG" != "$TOML_VERSION" ]; then
+            echo "::error::Tag $TAG does not match pyproject.toml version $TOML_VERSION"
+            exit 1
+          fi
+
       - run: uv build
 
       - uses: pypa/gh-action-pypi-publish@release/v1

{code_atlas_mcp-0.2.0.dev1 → code_atlas_mcp-0.4.0}/.gitignore

@@ -66,6 +66,9 @@ data/tei/
 # Local config overrides
 atlas.local.toml
 
+# Generated knowledge-vault landing page (atlas dream) — derived, not authored
+docs/HOME.md
+
 # Environment files (secrets) — .env.dist is tracked
 .env
 .env.*

{code_atlas_mcp-0.2.0.dev1 → code_atlas_mcp-0.4.0}/CHANGELOG.md

@@ -2,6 +2,180 @@
 
 <!-- version list -->
 
+## v0.4.0 (2026-07-14)
+
+### Bug Fixes
+
+- **analysis**: Derive cross-package coupling from real package names
+  ([`655de39`](https://github.com/SerPeter/code-atlas/commit/655de398f32d30a7f198d3cc6827fd537792fc10))
+
+- **graph**: Detect body-only edits and resolve cross-file references
+  ([`3e417af`](https://github.com/SerPeter/code-atlas/commit/3e417afa92cd6f8a4a4c5176906dc82f1ef4cfcd))
+
+- **graph**: Preserve cross-file edges on delete and stop BM25 crashes
+  ([`6f30442`](https://github.com/SerPeter/code-atlas/commit/6f30442463377a06d4058118b57dce3cc1ad9c45))
+
+- **indexing**: Load nested gitignores for the watcher, honor monorepo scope
+  ([`ab3128b`](https://github.com/SerPeter/code-atlas/commit/ab3128ba06d78f291f947963edcd7b72e95ac830))
+
+- **indexing**: Make the event pipeline durable and path-consistent
+  ([`a8b0246`](https://github.com/SerPeter/code-atlas/commit/a8b02460043e3aaeab38513bfeb1bd47128f1334))
+
+- **indexing**: Scope monorepo indexing and gate empty-scan deletion
+  ([`ba1405d`](https://github.com/SerPeter/code-atlas/commit/ba1405d5d7de281341522cc056c7f96e9bddd7a8))
+
+- **indexing**: Withhold file hashes until deferred edges resolve
+  ([`dcdeecc`](https://github.com/SerPeter/code-atlas/commit/dcdeecc64137589e13765e6c7965cecb273b8f41))
+
+- **infra**: Enable AOF persistence for the production Valkey event bus
+  ([`3751c20`](https://github.com/SerPeter/code-atlas/commit/3751c20b81ce9ef24d64d044a03c0587ab445337))
+
+- **parser**: Attach cross-file Go methods via receiver type name
+  ([`b24b0d3`](https://github.com/SerPeter/code-atlas/commit/b24b0d353cb63de930468af5fd6a415b515808f3))
+
+- **parser**: Capture multi-name Go var/const specs and exclude generics from USES_TYPE
+  ([`81ca9d2`](https://github.com/SerPeter/code-atlas/commit/81ca9d2cf38ea947709e9b84276be1dc3bb1c9d4))
+
+- **parser**: Disambiguate duplicate Markdown headings and fix line_end
+  ([`a5c89c4`](https://github.com/SerPeter/code-atlas/commit/a5c89c4526f6373fb73b70712201b0a985f5fffb))
+
+- **parser**: Disambiguate JVM overload uids with parameter signatures
+  ([`e0fe188`](https://github.com/SerPeter/code-atlas/commit/e0fe188875dd492f7b49b1ee81a05a8b364cdf7c))
+
+- **parser**: Extract C++ operator overloads and out-of-line nested definitions
+  ([`5acea2a`](https://github.com/SerPeter/code-atlas/commit/5acea2a38745535a407558233b8ef01fdd96df8b))
+
+- **parser**: Extract C++ templates, prototypes, and out-of-line methods
+  ([`805ca54`](https://github.com/SerPeter/code-atlas/commit/805ca5447aed09e659630fed941a46e6354e1ef5))
+
+- **parser**: Extract Ruby inline-visibility methods
+  ([`b76f0bd`](https://github.com/SerPeter/code-atlas/commit/b76f0bdd87eb40ce8235c5c2f8b4a865f1696f58))
+
+- **parser**: Extract TS interface heritage, re-exports, and decorators
+  ([`1278201`](https://github.com/SerPeter/code-atlas/commit/1278201731ea058fb5f4a5f5f06e691b1c4a0eb9))
+
+- **parser**: Fold JVM namespaces into qualified names, resolve Java imports
+  ([`6199504`](https://github.com/SerPeter/code-atlas/commit/619950430c739301517154a69f2882cb539a587b))
+
+- **parser**: Parse .tsx files with the TSX grammar
+  ([`82a3db3`](https://github.com/SerPeter/code-atlas/commit/82a3db3acc9ddf458af4afbd3a1e080f988c158b))
+
+- **parser**: Preserve compact Ruby class-path names for INHERITS matching
+  ([`6449af5`](https://github.com/SerPeter/code-atlas/commit/6449af5a97ce6ab089f4b1cba688bb301b96ecd3))
+
+- **parser**: Resolve nested-class names and stop plugin-load lockout
+  ([`136e78c`](https://github.com/SerPeter/code-atlas/commit/136e78cd31e47d7c3c1be40b6df0292b1e1e7872))
+
+- **parser**: Scope Rust associated types to their impl block
+  ([`7cdc801`](https://github.com/SerPeter/code-atlas/commit/7cdc8011edae9b06c8f91fd4c17aaaacf5359011))
+
+- **parser**: Walk braced PHP namespaces and namespace-qualify entity names
+  ([`16efc39`](https://github.com/SerPeter/code-atlas/commit/16efc39b44bed00376d2a37c7a0112d7a3ca1cff))
+
+- **parser**: Walk inline Rust modules and cross-file impl parents
+  ([`a4c1d5e`](https://github.com/SerPeter/code-atlas/commit/a4c1d5eefcf4f4bfce63f0f350aec09d396b274e))
+
+- **search**: Stop suppressing vector search and silently emptying scope filters
+  ([`d78f86d`](https://github.com/SerPeter/code-atlas/commit/d78f86d136fd3f2122991d224b23f32448bd0dc0))
+
+- **server**: Correct diagram scoping, cycle detection, and node-cap edge handling
+  ([`d919f1e`](https://github.com/SerPeter/code-atlas/commit/d919f1e958f8be02220f0fd97c10b6eff6626b8d))
+
+- **server**: Fix cypher_query serialization and scope defaults for monorepos
+  ([`e5a8c7f`](https://github.com/SerPeter/code-atlas/commit/e5a8c7fc5019545c3c63bb7955dfdc998850d215))
+
+- **server**: Validate label params and surface pipeline health honestly
+  ([`5c35cbe`](https://github.com/SerPeter/code-atlas/commit/5c35cbea4f54a87d425030b371e3272bc8222601))
+
+- **settings**: Resolve atlas.toml against project_root, not cwd
+  ([`3c3b21c`](https://github.com/SerPeter/code-atlas/commit/3c3b21cdbf3f1d6e284465e5d157b2f532a34a30))
+
+- **settings**: Scope nested config sections to prefixed env vars
+  ([`f1aa4d9`](https://github.com/SerPeter/code-atlas/commit/f1aa4d9d0f809cc5913969c501d3572edd2bfda1))
+
+### Chores
+
+- Drop unused type-ignore directives
+  ([`0f7b85a`](https://github.com/SerPeter/code-atlas/commit/0f7b85ab3cb8a037093f1eae3c1ab89f92000319))
+
+- Sync uv.lock to the 0.3.1 version bump
+  ([`5b5189b`](https://github.com/SerPeter/code-atlas/commit/5b5189bd4af3055fff224bf616a2d29e9a7e4034))
+
+### Documentation
+
+- Add atlas dream to the CLI commands list
+  ([`b52d921`](https://github.com/SerPeter/code-atlas/commit/b52d921b107a886790fdffea00ec6e80d79c1342))
+
+- **adr**: Record cross-file resolution, pipeline durability, and test isolation decisions
+  ([`ada1ce1`](https://github.com/SerPeter/code-atlas/commit/ada1ce113111ef781b4cf7cbebea64de1f6b053f))
+
+- **adr**: Record the Note vault schema decision
+  ([`d93215c`](https://github.com/SerPeter/code-atlas/commit/d93215cc383e39732648cd4a966978be7095085b))
+
+### Features
+
+- **knowledge**: Add anchors + staleness resolution (Phase 3)
+  ([`c22ebd6`](https://github.com/SerPeter/code-atlas/commit/c22ebd61677f098b5a3c727d9fb0640f98ade272))
+
+- **knowledge**: Add dream-mode deterministic report (Phase 4)
+  ([`47c4a6d`](https://github.com/SerPeter/code-atlas/commit/47c4a6d86851f73c9abcd4535f8e42186cd546ff))
+
+- **knowledge**: Add Note vault foundations to the code graph (Phase 1)
+  ([`1706a0a`](https://github.com/SerPeter/code-atlas/commit/1706a0acfeae3958b82f3f004390a0331f0dea4b))
+
+- **knowledge**: Live global vault + polish (Phase 5)
+  ([`9b40136`](https://github.com/SerPeter/code-atlas/commit/9b40136091ef147deac443c55b06e2e478facac4))
+
+- **knowledge**: Poll and index extra vaults from the daemon (Phase 2)
+  ([`ef750bb`](https://github.com/SerPeter/code-atlas/commit/ef750bbb46d658c0232b4a8f32734169d915969e))
+
+### Testing
+
+- **integration**: Isolate tests from production databases
+  ([`a1bdecd`](https://github.com/SerPeter/code-atlas/commit/a1bdecd4bdfa8225e894442bd62c0facc79e6654))
+
+
+## v0.3.1 (2026-03-07)
+
+### Bug Fixes
+
+- **ci**: Use GitHub App token to push to protected main branch
+  ([`13b46b9`](https://github.com/SerPeter/code-atlas/commit/13b46b972f900e4d1170b9c42af0c40d9c0cf35c))
+
+
+## v0.3.0 (2026-03-04)
+
+### Features
+
+- **indexing**: Add file hash gate to skip unchanged files
+  ([#9](https://github.com/SerPeter/code-atlas/pull/9))
+- **indexing**: Add per-file cooldown for daemon mode
+  ([#9](https://github.com/SerPeter/code-atlas/pull/9))
+
+### Performance Improvements
+
+- Reduce RTTs across indexing and query pipelines
+  ([`cf3a519`](https://github.com/SerPeter/code-atlas/commit/cf3a519))
+
+### Refactoring
+
+- **indexing**: Eliminate Tier 1 consumer, simplify to two-tier pipeline
+  ([#9](https://github.com/SerPeter/code-atlas/pull/9))
+- **indexing**: Rename Tier 2/3 to AST/Embed stage across code and docs
+  ([#9](https://github.com/SerPeter/code-atlas/pull/9))
+
+### Bug Fixes
+
+- **ci**: Remove detached HEAD checkout in release workflow
+
+## v0.2.0-dev.2 (2026-02-23)
+
+### Bug Fixes
+
+- **indexing**: Label-constrained queries, PEL reclaim, and drain progress
+  ([`9ec9150`](https://github.com/SerPeter/code-atlas/commit/9ec915088058b05396a506ca23bfd0a0492d67f6))
+
+
 ## v0.2.0-dev.1 (2026-02-23)
 
 ### Bug Fixes
@@ -13,14 +187,15 @@
 - **ci**: Resolve ty check failures with --all-extras in CI
   ([`3f74816`](https://github.com/SerPeter/code-atlas/commit/3f7481635091d2d676aed75c3fbcaa5db4332242))
 
-- **consumers**: Group batches by project in Tier1/Tier2
+- **consumers**: Group batches by project in AST/Embed consumers
   ([#2](https://github.com/SerPeter/code-atlas/pull/2),
   [`5107b24`](https://github.com/SerPeter/code-atlas/commit/5107b24a7dfbcb44cadc7917f632ae6a9743c057))
 
 ### Build System
 
 - **release**: Add python-semantic-release v10 for automated releases
-  ([`044e9a3`](https://github.com/SerPeter/code-atlas/commit/044e9a31c5f7673a4be45a6519860d247e1b3cb1))
+  ([#5](https://github.com/SerPeter/code-atlas/pull/5),
+  [`de9c06d`](https://github.com/SerPeter/code-atlas/commit/de9c06d8c1e1744a92ff7ef160e06f859b5aa768))
 
 ### Features
 
@@ -181,7 +356,7 @@
 - **docs**: Add markdown parser with tree-sitter-markdown
   ([`e8d372c`](https://github.com/SerPeter/code-atlas/commit/e8d372c162652d6d73d1f66da5e14a61fcb2136a))
 
-- **embeddings**: Add EmbedClient with litellm routing and Tier 3 pipeline
+- **embeddings**: Add EmbedClient with litellm routing and embed pipeline
   ([`ad7c972`](https://github.com/SerPeter/code-atlas/commit/ad7c9726f2e48fdb8746b50547089c5c483bcb75))
 
 - **embeddings**: Add three-tier embedding cache with Valkey backend
@@ -232,7 +407,7 @@
 - **naming**: Worktree-aware naming and monorepo sub-project prefixing
   ([`2acdfb3`](https://github.com/SerPeter/code-atlas/commit/2acdfb33ba4b486f966272a01cf8a37f670661f6))
 
-- **parser**: Add py-tree-sitter parser, implement Tier 2 pipeline, drop Rust
+- **parser**: Add py-tree-sitter parser, implement AST pipeline, drop Rust
   ([`d56e7d2`](https://github.com/SerPeter/code-atlas/commit/d56e7d2a686ec279a52d85bbc4903f4d85f51a4e))
 
 - **parsing**: Add multi-language support (10 languages, 7 modules)

{code_atlas_mcp-0.2.0.dev1 → code_atlas_mcp-0.4.0}/CLAUDE.md

@@ -18,7 +18,7 @@ uv sync --group dev              # Include dev dependencies
 # Run tests
 uv run pytest                    # All tests
 uv run pytest -m "not slow"      # Skip slow tests
-uv run pytest -m integration     # Integration tests only (requires Memgraph)
+uv run pytest -m integration     # Integration tests only (requires Docker — testcontainers by default, see Testing)
 uv run pytest tests/test_foo.py::test_bar  # Single test
 
 # Lint and format
@@ -32,7 +32,8 @@ uv run pre-commit install        # Install hooks
 uv run pre-commit run --all-files  # Run all hooks manually
 
 # Infrastructure
-docker compose up -d             # Start Memgraph + Valkey
+docker compose up -d             # Start Memgraph + Valkey (production index: 7687/6379)
+docker compose --profile test up -d  # Optional integration-test fast path (memgraph-test :7688, valkey-test :6380, see Testing)
 docker compose --profile tei up -d  # Include local embeddings (TEI)
 docker compose down              # Stop services
 
@@ -42,6 +43,8 @@ atlas search "query"             # Hybrid search
 atlas status                     # Check index status
 atlas mcp                        # Start MCP server
 atlas daemon start               # Start indexing daemon (watcher + pipeline)
+atlas dream                      # Knowledge-vault lint report (inbox, orphans, dangling links, duplicates) + docs/HOME.md
+atlas project rm <name>          # Delete a project's graph data (e.g. a stale worktree project)
 ```
 
 ## Architecture
@@ -51,7 +54,7 @@ src/code_atlas/
 ├── __init__.py          # __version__ only
 ├── schema.py            # Graph schema (labels, relationships, DDL generators)
 ├── settings.py          # Pydantic configuration (atlas.toml + env vars)
-├── events.py            # Event types (FileChanged, ASTDirty, EmbedDirty) + Valkey Streams EventBus
+├── events.py            # Event types (FileChanged, EmbedDirty) + Valkey Streams EventBus
 ├── telemetry.py         # OpenTelemetry integration
 ├── cli.py               # Typer CLI entrypoint (index, search, status, mcp, daemon commands)
 │
@@ -69,7 +72,7 @@ src/code_atlas/
 │
 ├── indexing/
 │   ├── orchestrator.py  # Full-index, monorepo detection, staleness checking
-│   ├── consumers.py     # Tier 1/2/3 event consumers (batch-pull pattern)
+│   ├── consumers.py     # AST + Embed event consumers (batch-pull pattern)
 │   ├── watcher.py       # Filesystem watcher (watchfiles + hybrid debounce)
 │   └── daemon.py        # Daemon lifecycle manager (watcher + pipeline)
 │
@@ -78,12 +81,14 @@ src/code_atlas/
     └── health.py        # Infrastructure health checks + diagnostics
 ```
 
-**Event Pipeline:** File Watcher → Valkey Streams → Tier 1 (graph metadata) → Tier 2 (AST diff + gate) → Tier 3 (embeddings) → Memgraph
+**Event Pipeline:** File Watcher → Valkey Streams → AST stage (hash gate + parse + diff) → Embed stage (embeddings) → Memgraph
 
 **Query Pipeline:** MCP Server → Query Router → [Graph Search | Vector Search | BM25 Search] → RRF Fusion → Results
 
 **Deployment:** Daemon (`atlas daemon start`) for indexing + MCP (`atlas mcp`) per agent session, decoupled via Valkey + Memgraph
 
+**Event model:** Events are atomic — one logical change per event (one file per FileChanged, one entity per EmbedDirty). Never bundle lists of work items into a single event; use `EventBus.publish_many()` for network-efficient batch publishing. The consumer's `max_batch_size` must directly control work volume, not just message count.
+
 **Infrastructure:** Memgraph (graph DB, port 7687), TEI (embeddings, port 8080), Valkey (event bus, port 6379)
 
 ## Code Style
@@ -126,6 +131,7 @@ src/code_atlas/
 - **High gear (default):** Integration tests exercising full workflows and public APIs
 - **Low gear (selective):** Unit tests only for complex algorithms or edge cases unreachable via integration
 - Don't test every function. Test system behavior.
+- Integration tests start session-scoped testcontainers on random ports by default (Docker required; skip if unavailable). Fast path: `docker compose --profile test up -d`, then `ATLAS_TEST_MEMGRAPH_PORT=7688 ATLAS_TEST_VALKEY_PORT=6380 uv run pytest -m integration` — the env vars point tests at any isolated stack (e.g. CI service containers). Never connects to the production Memgraph/Valkey on 7687/6379. A conftest guard refuses to wipe any Memgraph containing project data not prefixed `test`/`bench`; `ATLAS_TEST_DB=1` bypasses it for known-disposable instances only.
 
 ## Commits