context-rail 3.0.2__tar.gz → 3.0.4__tar.gz

context_rail-3.0.4/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Son Hoanh
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

{context_rail-3.0.2/src/context_rail.egg-info → context_rail-3.0.4}/PKG-INFO

@@ -1,15 +1,29 @@
 Metadata-Version: 2.4
 Name: context-rail
-Version: 3.0.2
+Version: 3.0.4
 Summary: An agentic project workspace: files are the truth, agents work through one write path
+License-Expression: MIT
 Project-URL: Repository, https://github.com/nguyenhoanhson797/Context-Rail
+Keywords: agents,context,project-management,markdown,sqlite,workflow,mcp,cli
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown
+License-File: LICENSE
 Requires-Dist: PyYAML>=6.0
 Requires-Dist: mcp<3,>=2.3.0
 Provides-Extra: test
 Requires-Dist: pytest>=8; extra == "test"
 Requires-Dist: pytest-timeout>=2.3; extra == "test"
+Dynamic: license-file
 
 # Context Rail v3
 
@@ -30,19 +44,28 @@ the package ships the CLI, the MCP server and a loopback UI over one tool table.
 
 ## Install
 
-```powershell
-py -m pip install context-rail
+```bash
+python -m pip install context-rail
 ```
 
-On macOS or Linux use `python` instead of `py`. This installs two equivalent
-entry points, `rail` and `ctx-rail`; every example below also works as
-`py -m context_rail`.
+On Windows use `py` instead of `python` (`py -m pip install context-rail`); that
+is the only difference, every command below works the same in both shells.
+This installs two equivalent entry points, `rail` and `ctx-rail`; every example
+below also works as `python -m context_rail`.
+
+`python -m pip` installs into the interpreter that will run the tool. Bare
+`pip install context-rail` is the same command whenever `pip` on your `PATH`
+belongs to that interpreter; on a machine with several Pythons, the module form
+is the one that cannot pick the wrong one.
+
+To keep the tool out of your project environments, `pipx install context-rail`
+works too.
 
 ## Quick start
 
 From the project you want to manage:
 
-```powershell
+```bash
 rail init --pack basic            # writes ./context-rail/
 rail add goal "Ship v3"
 rail add task "Write the first task" --rel part_of=ship-v3
@@ -67,7 +90,7 @@ Notes that are easy to trip over:
 `rail init --pack basic` writes only the skeleton (`rail.yaml`, `.gitignore`,
 empty folders). For a picture, the profiles and the agent files:
 
-```powershell
+```bash
 rail init --interview                                   # or --answers-file FILE --confirm "APPROVE charter"
 rail approve charter --confirm "APPROVE charter" --confirmer "Ada"
 rail adapters sync                                      # AGENTS.md + .agents/skills/
@@ -84,7 +107,7 @@ rail sync --install                                     # optional: git merge dr
 
 Useful read commands:
 
-```powershell
+```bash
 rail status            # counts, executable items, blockers, stuck items
 rail digest            # activity, waits, decisions, quality, usage
 rail cto brief         # the bounded brief a CTO session starts from
@@ -96,19 +119,23 @@ rail ci                # doctor issues plus the approval and done gates
 
 ## Skills
 
-Four operator skills under `skills/` teach an agent the rail itself, so it does
-not have to read the engine to answer "what can this do?" or "how do we change
-the workflow?":
+Four operator skills under `skills/` in the repository checkout teach an agent
+the rail itself, so it does not have to read the engine to answer "what can this
+do?" or "how do we change the workflow?":
 
 - `context-rail-v3` — mental model, quick start, command map
 - `context-rail-v3-workflows` — packs, statuses, transitions, BA/PO steps
 - `context-rail-v3-loop` — dispatch → verify → review → merge, refusals
 - `context-rail-v3-ops` — doctor, reindex, recovery, budgets, test lanes
 
-Install them into your agent's skill directory:
+Install them into your agent's skill directory from a checkout:
+
+```bash
+cp -r skills/context-rail-v3* ~/.agents/skills/          # macOS, Linux
+```
 
 ```powershell
-Copy-Item -Recurse skills\context-rail-v3* "$HOME\.agents\skills\"
+Copy-Item -Recurse skills\context-rail-v3* "$HOME\.agents\skills\"   # Windows
 ```
 
 See `skills/README.md` for the difference from the pack skills that are
@@ -120,13 +147,13 @@ Work that launches an agent, or changes the picture's meaning, is gated. The
 brief reports the current state; nothing executes until a human approves the
 charter:
 
-```powershell
+```bash
 rail approve charter --confirm "APPROVE charter" --confirmer "Ada"
 ```
 
 Once the charter is approved, the loop uses isolated git worktrees:
 
-```powershell
+```bash
 rail dispatch <item> --expect <etag> --key <idempotency-key>
 rail review   <item> --run <run-id> --key <idempotency-key>
 rail merge    <item> --run <run-id> --expect <etag> --key <idempotency-key>
@@ -168,20 +195,20 @@ are all generated from one tool table (`src/context_rail/gateway/tools.py`).
 
 From a checkout, install it in editable mode with the test dependencies:
 
-```powershell
-py -m pip install -e .[test]
+```bash
+python -m pip install -e .[test]
 ```
 
-Then run one lane at a time:
+Then run one lane at a time (on Windows use `py -m pytest`):
 
-```powershell
-py -m pytest -q -m "not slow and not perf"          # functional lane
-py -m pytest -q -m "perf and not slow" -p no:xdist  # speed budgets, serial
-py -m pytest -q -m slow -p no:xdist                 # 10k-document stores
+```bash
+python -m pytest -q -m "not slow and not perf"          # functional lane
+python -m pytest -q -m "perf and not slow" -p no:xdist  # speed budgets, serial
+python -m pytest -q -m slow -p no:xdist                 # 10k-document stores
 ```
 
 The speed budgets are fixed: a change that needs more time is a bug in the
-change, never a reason to raise the budget. `py -m pytest -q` on its own also
+change, never a reason to raise the budget. A bare `pytest -q` on its own also
 collects the perf tests, so use the lanes above when the machine is busy.
 
 ## Layout
@@ -203,13 +230,16 @@ src/context_rail/
 
 ## Design
 
-- Visual guide (diagrams and from-zero-to-daily walkthrough):
-  `docs/context-rail-v3-guide.html`
-- Interactive engine graph (primitives, relations, one write path and the agent loop):
-  `docs/context-rail-v3-engine-graph.html`
-- Spec: `docs/superpowers/specs/2026-10-28-context-rail-v3-design.md`
-- Architecture: `library/arch-stage-3.md`
-- Working rules and engine invariants: `AGENTS.md`
+- Visual guide (diagrams and a from-zero-to-daily walkthrough):
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/docs/context-rail-v3-guide.html>
+- Interactive engine graph (primitives, relations, one write path and the agent
+  loop):
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/docs/context-rail-v3-engine-graph.html>
+- Working rules and engine invariants:
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/AGENTS.md>
+
+The design spec, the stage plans and the architecture records are working
+material and stay out of the published repository.
 
 Engine invariants worth knowing before changing code:
 
@@ -218,3 +248,8 @@ Engine invariants worth knowing before changing code:
 - `context_rail.kernel` is pure: no file, git, process or index access.
 - Unknown manifest schema, pack version, key or policy field is refused, never
   guessed.
+
+## License
+
+MIT. See the `LICENSE` file, or
+<https://github.com/nguyenhoanhson797/Context-Rail/blob/master/LICENSE>.

context_rail-3.0.2/PKG-INFO → context_rail-3.0.4/README.md

@@ -1,16 +1,3 @@
-Metadata-Version: 2.4
-Name: context-rail
-Version: 3.0.2
-Summary: An agentic project workspace: files are the truth, agents work through one write path
-Project-URL: Repository, https://github.com/nguyenhoanhson797/Context-Rail
-Requires-Python: >=3.11
-Description-Content-Type: text/markdown
-Requires-Dist: PyYAML>=6.0
-Requires-Dist: mcp<3,>=2.3.0
-Provides-Extra: test
-Requires-Dist: pytest>=8; extra == "test"
-Requires-Dist: pytest-timeout>=2.3; extra == "test"
-
 # Context Rail v3
 
 An agentic project workspace: the files are the truth, a rebuildable SQLite
@@ -30,19 +17,28 @@ the package ships the CLI, the MCP server and a loopback UI over one tool table.
 
 ## Install
 
-```powershell
-py -m pip install context-rail
+```bash
+python -m pip install context-rail
 ```
 
-On macOS or Linux use `python` instead of `py`. This installs two equivalent
-entry points, `rail` and `ctx-rail`; every example below also works as
-`py -m context_rail`.
+On Windows use `py` instead of `python` (`py -m pip install context-rail`); that
+is the only difference, every command below works the same in both shells.
+This installs two equivalent entry points, `rail` and `ctx-rail`; every example
+below also works as `python -m context_rail`.
+
+`python -m pip` installs into the interpreter that will run the tool. Bare
+`pip install context-rail` is the same command whenever `pip` on your `PATH`
+belongs to that interpreter; on a machine with several Pythons, the module form
+is the one that cannot pick the wrong one.
+
+To keep the tool out of your project environments, `pipx install context-rail`
+works too.
 
 ## Quick start
 
 From the project you want to manage:
 
-```powershell
+```bash
 rail init --pack basic            # writes ./context-rail/
 rail add goal "Ship v3"
 rail add task "Write the first task" --rel part_of=ship-v3
@@ -67,7 +63,7 @@ Notes that are easy to trip over:
 `rail init --pack basic` writes only the skeleton (`rail.yaml`, `.gitignore`,
 empty folders). For a picture, the profiles and the agent files:
 
-```powershell
+```bash
 rail init --interview                                   # or --answers-file FILE --confirm "APPROVE charter"
 rail approve charter --confirm "APPROVE charter" --confirmer "Ada"
 rail adapters sync                                      # AGENTS.md + .agents/skills/
@@ -84,7 +80,7 @@ rail sync --install                                     # optional: git merge dr
 
 Useful read commands:
 
-```powershell
+```bash
 rail status            # counts, executable items, blockers, stuck items
 rail digest            # activity, waits, decisions, quality, usage
 rail cto brief         # the bounded brief a CTO session starts from
@@ -96,19 +92,23 @@ rail ci                # doctor issues plus the approval and done gates
 
 ## Skills
 
-Four operator skills under `skills/` teach an agent the rail itself, so it does
-not have to read the engine to answer "what can this do?" or "how do we change
-the workflow?":
+Four operator skills under `skills/` in the repository checkout teach an agent
+the rail itself, so it does not have to read the engine to answer "what can this
+do?" or "how do we change the workflow?":
 
 - `context-rail-v3` — mental model, quick start, command map
 - `context-rail-v3-workflows` — packs, statuses, transitions, BA/PO steps
 - `context-rail-v3-loop` — dispatch → verify → review → merge, refusals
 - `context-rail-v3-ops` — doctor, reindex, recovery, budgets, test lanes
 
-Install them into your agent's skill directory:
+Install them into your agent's skill directory from a checkout:
+
+```bash
+cp -r skills/context-rail-v3* ~/.agents/skills/          # macOS, Linux
+```
 
 ```powershell
-Copy-Item -Recurse skills\context-rail-v3* "$HOME\.agents\skills\"
+Copy-Item -Recurse skills\context-rail-v3* "$HOME\.agents\skills\"   # Windows
 ```
 
 See `skills/README.md` for the difference from the pack skills that are
@@ -120,13 +120,13 @@ Work that launches an agent, or changes the picture's meaning, is gated. The
 brief reports the current state; nothing executes until a human approves the
 charter:
 
-```powershell
+```bash
 rail approve charter --confirm "APPROVE charter" --confirmer "Ada"
 ```
 
 Once the charter is approved, the loop uses isolated git worktrees:
 
-```powershell
+```bash
 rail dispatch <item> --expect <etag> --key <idempotency-key>
 rail review   <item> --run <run-id> --key <idempotency-key>
 rail merge    <item> --run <run-id> --expect <etag> --key <idempotency-key>
@@ -168,20 +168,20 @@ are all generated from one tool table (`src/context_rail/gateway/tools.py`).
 
 From a checkout, install it in editable mode with the test dependencies:
 
-```powershell
-py -m pip install -e .[test]
+```bash
+python -m pip install -e .[test]
 ```
 
-Then run one lane at a time:
+Then run one lane at a time (on Windows use `py -m pytest`):
 
-```powershell
-py -m pytest -q -m "not slow and not perf"          # functional lane
-py -m pytest -q -m "perf and not slow" -p no:xdist  # speed budgets, serial
-py -m pytest -q -m slow -p no:xdist                 # 10k-document stores
+```bash
+python -m pytest -q -m "not slow and not perf"          # functional lane
+python -m pytest -q -m "perf and not slow" -p no:xdist  # speed budgets, serial
+python -m pytest -q -m slow -p no:xdist                 # 10k-document stores
 ```
 
 The speed budgets are fixed: a change that needs more time is a bug in the
-change, never a reason to raise the budget. `py -m pytest -q` on its own also
+change, never a reason to raise the budget. A bare `pytest -q` on its own also
 collects the perf tests, so use the lanes above when the machine is busy.
 
 ## Layout
@@ -203,13 +203,16 @@ src/context_rail/
 
 ## Design
 
-- Visual guide (diagrams and from-zero-to-daily walkthrough):
-  `docs/context-rail-v3-guide.html`
-- Interactive engine graph (primitives, relations, one write path and the agent loop):
-  `docs/context-rail-v3-engine-graph.html`
-- Spec: `docs/superpowers/specs/2026-10-28-context-rail-v3-design.md`
-- Architecture: `library/arch-stage-3.md`
-- Working rules and engine invariants: `AGENTS.md`
+- Visual guide (diagrams and a from-zero-to-daily walkthrough):
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/docs/context-rail-v3-guide.html>
+- Interactive engine graph (primitives, relations, one write path and the agent
+  loop):
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/docs/context-rail-v3-engine-graph.html>
+- Working rules and engine invariants:
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/AGENTS.md>
+
+The design spec, the stage plans and the architecture records are working
+material and stay out of the published repository.
 
 Engine invariants worth knowing before changing code:
 
@@ -218,3 +221,8 @@ Engine invariants worth knowing before changing code:
 - `context_rail.kernel` is pure: no file, git, process or index access.
 - Unknown manifest schema, pack version, key or policy field is refused, never
   guessed.
+
+## License
+
+MIT. See the `LICENSE` file, or
+<https://github.com/nguyenhoanhson797/Context-Rail/blob/master/LICENSE>.

{context_rail-3.0.2 → context_rail-3.0.4}/pyproject.toml

@@ -4,11 +4,26 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "context-rail"
-version = "3.0.2"
+version = "3.0.4"
 description = "An agentic project workspace: files are the truth, agents work through one write path"
 readme = "README.md"
+license = "MIT"
+license-files = ["LICENSE"]
 requires-python = ">=3.11"
 dependencies = ["PyYAML>=6.0", "mcp>=2.3.0,<3"]
+keywords = ["agents", "context", "project-management", "markdown", "sqlite", "workflow", "mcp", "cli"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Environment :: Console",
+    "Intended Audience :: Developers",
+    "Operating System :: OS Independent",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3 :: Only",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Topic :: Software Development",
+]
 
 [project.optional-dependencies]
 test = ["pytest>=8", "pytest-timeout>=2.3"]

context_rail-3.0.2/README.md → context_rail-3.0.4/src/context_rail.egg-info/PKG-INFO

@@ -1,3 +1,30 @@
+Metadata-Version: 2.4
+Name: context-rail
+Version: 3.0.4
+Summary: An agentic project workspace: files are the truth, agents work through one write path
+License-Expression: MIT
+Project-URL: Repository, https://github.com/nguyenhoanhson797/Context-Rail
+Keywords: agents,context,project-management,markdown,sqlite,workflow,mcp,cli
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: PyYAML>=6.0
+Requires-Dist: mcp<3,>=2.3.0
+Provides-Extra: test
+Requires-Dist: pytest>=8; extra == "test"
+Requires-Dist: pytest-timeout>=2.3; extra == "test"
+Dynamic: license-file
+
 # Context Rail v3
 
 An agentic project workspace: the files are the truth, a rebuildable SQLite
@@ -17,19 +44,28 @@ the package ships the CLI, the MCP server and a loopback UI over one tool table.
 
 ## Install
 
-```powershell
-py -m pip install context-rail
+```bash
+python -m pip install context-rail
 ```
 
-On macOS or Linux use `python` instead of `py`. This installs two equivalent
-entry points, `rail` and `ctx-rail`; every example below also works as
-`py -m context_rail`.
+On Windows use `py` instead of `python` (`py -m pip install context-rail`); that
+is the only difference, every command below works the same in both shells.
+This installs two equivalent entry points, `rail` and `ctx-rail`; every example
+below also works as `python -m context_rail`.
+
+`python -m pip` installs into the interpreter that will run the tool. Bare
+`pip install context-rail` is the same command whenever `pip` on your `PATH`
+belongs to that interpreter; on a machine with several Pythons, the module form
+is the one that cannot pick the wrong one.
+
+To keep the tool out of your project environments, `pipx install context-rail`
+works too.
 
 ## Quick start
 
 From the project you want to manage:
 
-```powershell
+```bash
 rail init --pack basic            # writes ./context-rail/
 rail add goal "Ship v3"
 rail add task "Write the first task" --rel part_of=ship-v3
@@ -54,7 +90,7 @@ Notes that are easy to trip over:
 `rail init --pack basic` writes only the skeleton (`rail.yaml`, `.gitignore`,
 empty folders). For a picture, the profiles and the agent files:
 
-```powershell
+```bash
 rail init --interview                                   # or --answers-file FILE --confirm "APPROVE charter"
 rail approve charter --confirm "APPROVE charter" --confirmer "Ada"
 rail adapters sync                                      # AGENTS.md + .agents/skills/
@@ -71,7 +107,7 @@ rail sync --install                                     # optional: git merge dr
 
 Useful read commands:
 
-```powershell
+```bash
 rail status            # counts, executable items, blockers, stuck items
 rail digest            # activity, waits, decisions, quality, usage
 rail cto brief         # the bounded brief a CTO session starts from
@@ -83,19 +119,23 @@ rail ci                # doctor issues plus the approval and done gates
 
 ## Skills
 
-Four operator skills under `skills/` teach an agent the rail itself, so it does
-not have to read the engine to answer "what can this do?" or "how do we change
-the workflow?":
+Four operator skills under `skills/` in the repository checkout teach an agent
+the rail itself, so it does not have to read the engine to answer "what can this
+do?" or "how do we change the workflow?":
 
 - `context-rail-v3` — mental model, quick start, command map
 - `context-rail-v3-workflows` — packs, statuses, transitions, BA/PO steps
 - `context-rail-v3-loop` — dispatch → verify → review → merge, refusals
 - `context-rail-v3-ops` — doctor, reindex, recovery, budgets, test lanes
 
-Install them into your agent's skill directory:
+Install them into your agent's skill directory from a checkout:
+
+```bash
+cp -r skills/context-rail-v3* ~/.agents/skills/          # macOS, Linux
+```
 
 ```powershell
-Copy-Item -Recurse skills\context-rail-v3* "$HOME\.agents\skills\"
+Copy-Item -Recurse skills\context-rail-v3* "$HOME\.agents\skills\"   # Windows
 ```
 
 See `skills/README.md` for the difference from the pack skills that are
@@ -107,13 +147,13 @@ Work that launches an agent, or changes the picture's meaning, is gated. The
 brief reports the current state; nothing executes until a human approves the
 charter:
 
-```powershell
+```bash
 rail approve charter --confirm "APPROVE charter" --confirmer "Ada"
 ```
 
 Once the charter is approved, the loop uses isolated git worktrees:
 
-```powershell
+```bash
 rail dispatch <item> --expect <etag> --key <idempotency-key>
 rail review   <item> --run <run-id> --key <idempotency-key>
 rail merge    <item> --run <run-id> --expect <etag> --key <idempotency-key>
@@ -155,20 +195,20 @@ are all generated from one tool table (`src/context_rail/gateway/tools.py`).
 
 From a checkout, install it in editable mode with the test dependencies:
 
-```powershell
-py -m pip install -e .[test]
+```bash
+python -m pip install -e .[test]
 ```
 
-Then run one lane at a time:
+Then run one lane at a time (on Windows use `py -m pytest`):
 
-```powershell
-py -m pytest -q -m "not slow and not perf"          # functional lane
-py -m pytest -q -m "perf and not slow" -p no:xdist  # speed budgets, serial
-py -m pytest -q -m slow -p no:xdist                 # 10k-document stores
+```bash
+python -m pytest -q -m "not slow and not perf"          # functional lane
+python -m pytest -q -m "perf and not slow" -p no:xdist  # speed budgets, serial
+python -m pytest -q -m slow -p no:xdist                 # 10k-document stores
 ```
 
 The speed budgets are fixed: a change that needs more time is a bug in the
-change, never a reason to raise the budget. `py -m pytest -q` on its own also
+change, never a reason to raise the budget. A bare `pytest -q` on its own also
 collects the perf tests, so use the lanes above when the machine is busy.
 
 ## Layout
@@ -190,13 +230,16 @@ src/context_rail/
 
 ## Design
 
-- Visual guide (diagrams and from-zero-to-daily walkthrough):
-  `docs/context-rail-v3-guide.html`
-- Interactive engine graph (primitives, relations, one write path and the agent loop):
-  `docs/context-rail-v3-engine-graph.html`
-- Spec: `docs/superpowers/specs/2026-10-28-context-rail-v3-design.md`
-- Architecture: `library/arch-stage-3.md`
-- Working rules and engine invariants: `AGENTS.md`
+- Visual guide (diagrams and a from-zero-to-daily walkthrough):
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/docs/context-rail-v3-guide.html>
+- Interactive engine graph (primitives, relations, one write path and the agent
+  loop):
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/docs/context-rail-v3-engine-graph.html>
+- Working rules and engine invariants:
+  <https://github.com/nguyenhoanhson797/Context-Rail/blob/master/AGENTS.md>
+
+The design spec, the stage plans and the architecture records are working
+material and stay out of the published repository.
 
 Engine invariants worth knowing before changing code:
 
@@ -205,3 +248,8 @@ Engine invariants worth knowing before changing code:
 - `context_rail.kernel` is pure: no file, git, process or index access.
 - Unknown manifest schema, pack version, key or policy field is refused, never
   guessed.
+
+## License
+
+MIT. See the `LICENSE` file, or
+<https://github.com/nguyenhoanhson797/Context-Rail/blob/master/LICENSE>.

{context_rail-3.0.2 → context_rail-3.0.4}/src/context_rail.egg-info/SOURCES.txt

@@ -1,3 +1,4 @@
+LICENSE
 README.md
 pyproject.toml
 src/context_rail/__init__.py