noah-code 0.2.3__tar.gz → 0.2.4__tar.gz

{noah_code-0.2.3 → noah_code-0.2.4}/.github/workflows/ci.yml

@@ -5,6 +5,9 @@ on:
   push:
     branches:
       - main
+    # Also run on version tags so the exact released ref is checked.
+    tags:
+      - "v*"
 
 permissions:
   contents: read
@@ -67,6 +70,31 @@ jobs:
       - name: Build distributions
         run: uv build --out-dir build-dist
 
+  integration:
+    name: Integration tests (network)
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    # Integration-marked tests hit the live network and can flake for reasons
+    # unrelated to the code; keep the signal visible without blocking PRs.
+    continue-on-error: true
+    steps:
+      - name: Check out source
+        uses: actions/checkout@v7
+        with:
+          persist-credentials: false
+
+      - name: Install uv and Python
+        uses: astral-sh/setup-uv@08807647e7069bb48b6ef5acd8ec9567f424441b # v8.1.0
+        with:
+          enable-cache: true
+          python-version: "3.12"
+
+      - name: Install locked dependencies
+        run: uv sync --locked --all-extras --dev
+
+      - name: Run integration tests
+        run: uv run pytest -m integration -q
+
   platform:
     name: ${{ matrix.name }} smoke test
     runs-on: ${{ matrix.runner }}
@@ -109,8 +137,18 @@ jobs:
           rg --version
 
       - name: Run tests
+        if: runner.os != 'macOS'
         run: uv run pytest tests -W error::pytest.PytestUnraisableExceptionWarning
 
+      - name: Run tests with coverage (macOS)
+        if: runner.os == 'macOS'
+        # Collect coverage on macOS too so mac-only paths are measured, but
+        # without a fail-under gate: per-platform coverage totals differ,
+        # and the 70% gate stays on the ubuntu quality job.
+        run: |
+          uv run pytest tests --cov=noah_code --cov-report=term \
+            -W error::pytest.PytestUnraisableExceptionWarning
+
       - name: Verify command entry point
         run: uv run noah --version
 

{noah_code-0.2.3 → noah_code-0.2.4}/.github/workflows/release.yml

@@ -53,6 +53,9 @@ jobs:
           sudo apt-get install -y ripgrep
           rg --version
 
+      - name: Type check
+        run: uv run mypy src/noah_code
+
       - name: Run release checks
         run: |
           uv run ruff check src tests

noah_code-0.2.4/.pre-commit-config.yaml

@@ -0,0 +1,21 @@
+# Pre-commit hooks mirroring the CI lint gate.
+# Install with: uv run pre-commit install  (or `pre-commit install` once installed)
+repos:
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.16.1
+    hooks:
+      - id: ruff-check
+        args: [--fix]
+
+  - repo: local
+    hooks:
+      - id: mypy
+        name: mypy
+        entry: uv run mypy
+        language: system
+        pass_filenames: false
+      - id: uv-lock-check
+        name: uv lock --check
+        entry: uv lock --check
+        language: system
+        pass_filenames: false

{noah_code-0.2.3 → noah_code-0.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: noah-code
-Version: 0.2.3
+Version: 0.2.4
 Summary: Noah Code terminal coding agent, built on NVIDIA OO Agents (NOOA)
 Project-URL: Homepage, https://github.com/skundu42/noah-code
 Project-URL: Documentation, https://github.com/skundu42/noah-code#readme
@@ -19,34 +19,36 @@ Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Software Development
 Requires-Python: <3.14,>=3.12
-Requires-Dist: click>=8.1.0
-Requires-Dist: litellm<1.92.0,>=1.84.0
+Requires-Dist: click<9.0.0,>=8.1.0
+Requires-Dist: litellm<1.99.0,>=1.96.0
 Requires-Dist: nooa-cli==0.0.9
 Requires-Dist: nooa==0.0.9
-Requires-Dist: packaging>=24.0
-Requires-Dist: pydantic>=2.5.0
-Requires-Dist: pyyaml>=6.0.0
-Requires-Dist: rich>=13.0.0
-Requires-Dist: textual>=1.0.0
+Requires-Dist: packaging<27.0,>=24.0
+Requires-Dist: pydantic<3.0.0,>=2.5.0
+Requires-Dist: pyyaml<7.0.0,>=6.0.0
+Requires-Dist: rich<16.0.0,>=13.0.0
+Requires-Dist: textual<9.0.0,>=1.0.0
 Provides-Extra: dev
-Requires-Dist: mypy>=1.13.0; extra == 'dev'
+Requires-Dist: mypy>=2.3.1; extra == 'dev'
 Requires-Dist: pytest-asyncio>=0.24.0; extra == 'dev'
 Requires-Dist: pytest-cov>=5.0.0; extra == 'dev'
+Requires-Dist: pytest-socket>=0.7.0; extra == 'dev'
 Requires-Dist: pytest>=8.0.0; extra == 'dev'
 Requires-Dist: ruff>=0.8.0; extra == 'dev'
+Requires-Dist: types-pyyaml>=6.0.12.20260815; extra == 'dev'
 Provides-Extra: mcp
 Requires-Dist: nooa[mcp]==0.0.9; extra == 'mcp'
 Provides-Extra: tracing
 Requires-Dist: nooa[tracing]==0.0.9; extra == 'tracing'
-Provides-Extra: tui
-Requires-Dist: textual>=1.0.0; extra == 'tui'
 Description-Content-Type: text/markdown
 
 <div align="center">
 
+<img src="https://raw.githubusercontent.com/skundu42/noah-code/main/docs/assets/noah-logo.svg" alt="Noah Code terminal wordmark" width="760">
+
 # Noah Code
 
-**A terminal coding agent built on NVIDIA's NOAA framework**
+**A durable, repository-aware terminal coding agent built on NVIDIA's NOOA framework**
 
 [![PyPI](https://img.shields.io/pypi/v/noah-code.svg)](https://pypi.org/project/noah-code/)
 [![CI](https://github.com/skundu42/noah-code/actions/workflows/ci.yml/badge.svg)](https://github.com/skundu42/noah-code/actions/workflows/ci.yml)
@@ -54,7 +56,7 @@ Description-Content-Type: text/markdown
 
 </div>
 
-![Noah Code handling a repository change in its terminal interface](docs/assets/noah-in-action.svg)
+![Noah Code handling a repository change in its terminal interface](https://raw.githubusercontent.com/skundu42/noah-code/main/docs/assets/noah-in-action.svg)
 
 <p align="center"><sub>The real Textual interface, captured from a deterministic Noah Code session.</sub></p>
 
@@ -93,10 +95,10 @@ configuration or session history.
   preimages, concurrent-change detection, immediate diagnostics, and rollback.
 - **Visible execution.** Stream bounded shell output while commands run, keep servers and watchers
   as managed background jobs, and revisit activity details with `F2`.
-- **Recoverable work.** Review staged and unstaged changes in `/diff`, then undo or redo journaled
-  file edits even after restarting Noah.
-- **Persistent sessions.** Resume repository-scoped conversations, todos, model choices, and
-  automatically compacted context without losing the full underlying tool results.
+- **Crash-safe work.** Durable file intents roll back interrupted workspace-tool writes, Git
+  checkpoints protect shell-driven changes, and interrupted model runs resume after restart.
+- **Persistent sessions.** Resume repository-scoped conversations, steering, todos, model choices,
+  budgets, background-job logs, and compacted context without losing full tool results.
 - **Token-efficient by construction.** Lean tool-output bounds with disk-backed recall, condensed
   subagent results, cache-stable request prefixes (volatile status arrives as appended events),
   and pointer-eviction compaction — measured live with `/tokens`.
@@ -104,6 +106,9 @@ configuration or session history.
   **plan** mode, with ordered `allow`, `ask`, and `deny` permission rules.
 - **Extensible workflows.** Add slash commands, opt-in skills, MCP servers, or markdown subagents;
   attach `@files` and images when the task needs more context.
+- **Long-running by design.** Provider retries and configurable fallback models handle transient
+  failures, while workspace leases, bounded artifacts, durable process ownership, and `/health`
+  keep unattended sessions observable.
 
 Noah follows repository instructions from `AGENTS.md`, `CLAUDE.md`, and
 `.noah-code/instructions.md`.
@@ -160,6 +165,10 @@ Type `/` to search the full command and configuration reference. The most common
 | --- | --- |
 | `Enter` | Send the current prompt or accept a selected suggestion |
 | `Shift+Enter` | Insert a newline |
+| Drag, then `Cmd+C` / `Ctrl+Shift+C` | Select and copy TUI text |
+| `Ctrl+Shift+C` with no selection | Copy the latest Noah reply |
+| `Ctrl+G` | Open the searchable skills picker |
+| `Ctrl+]` | Return to live transcript output |
 | `Tab` | Switch between build and plan mode |
 | `F2` | Open execution activity |
 | `F3` | Open paginated conversation history |
@@ -167,13 +176,32 @@ Type `/` to search the full command and configuration reference. The most common
 | `/theme` | Choose Atom One Dark, Noah Ocean, Graphite, or High Contrast |
 | `/diff` | Review staged and unstaged changes |
 | `/undo` / `/redo` | Traverse the persistent edit journal |
+| `/checkpoints` | List rolling Git worktree checkpoints |
+| `/health` | Inspect durable run, job, inbox, event, and artifact state |
 | `/tokens` | Inspect tokens, cache usage, prefix stability, model wait, and tool output |
 | `/efficiency` | Switch between `fast`, `balanced`, and `deep` budgets |
 
 On wide terminals, the side rail prioritizes the active operation, Git branch and change counts,
 session, model usage, update state, and plan. Git status is refreshed in the background at turn
 boundaries, so the animated working state stays responsive. The main pane remains centered on the
-Noah mark until the first prompt, then becomes the conversation and execution timeline.
+large Noah wordmark until the first prompt, then becomes the conversation and execution timeline.
+
+## Sessions and crash recovery
+
+Start a new session with `noah .`, resume the latest repository session with
+`noah --continue .`, or reopen an exact session with `noah --session SESSION_ID .`. Inside the TUI,
+use `Ctrl+O`, `/sessions`, or `/continue`.
+
+Noah stores conversational history and host runtime state separately. If the process stops during
+an active run, the next launch restores pending steering, usage and budget counters, durable job
+logs, and the original request. Incomplete workspace-tool writes are rolled back, verified orphan
+process groups are cleaned up, and non-interactive runs continue automatically. A request that was
+waiting for user input remains paused for the next user message.
+
+By default, only one Noah process may own a checkout at a time. Use `/worktree create` when
+independent agents need to work concurrently. See
+[Reliability and long-running sessions](docs/reliability.md) for the recovery model, provider retry
+controls, quotas, and operational limits.
 
 ## Models and providers
 
@@ -202,7 +230,8 @@ noah --model openai/MODEL --reasoning-effort high .
 ```
 
 See the [provider configuration guide](docs/configuration.md#bring-your-own-api-provider) for
-gateway-specific setup.
+gateway-specific setup. Provider request deadlines, exponential retry, and ordered fallback models
+are configured under `[reliability.retries]`.
 
 ## Updates
 
@@ -219,6 +248,7 @@ noah update
 
 - [Interactive interface and sessions](docs/interactive-reference.md)
 - [Configuration, modes, permissions, and updates](docs/configuration.md)
+- [Reliability and long-running sessions](docs/reliability.md)
 - [Generated-code security](docs/security.md)
 - [Custom commands, skills, MCP, and tracing](docs/extensions.md)
 - [Development, CI, and releases](docs/development.md)

{noah_code-0.2.3 → noah_code-0.2.4}/README.md

@@ -1,8 +1,10 @@
 <div align="center">
 
+<img src="https://raw.githubusercontent.com/skundu42/noah-code/main/docs/assets/noah-logo.svg" alt="Noah Code terminal wordmark" width="760">
+
 # Noah Code
 
-**A terminal coding agent built on NVIDIA's NOAA framework**
+**A durable, repository-aware terminal coding agent built on NVIDIA's NOOA framework**
 
 [![PyPI](https://img.shields.io/pypi/v/noah-code.svg)](https://pypi.org/project/noah-code/)
 [![CI](https://github.com/skundu42/noah-code/actions/workflows/ci.yml/badge.svg)](https://github.com/skundu42/noah-code/actions/workflows/ci.yml)
@@ -10,7 +12,7 @@
 
 </div>
 
-![Noah Code handling a repository change in its terminal interface](docs/assets/noah-in-action.svg)
+![Noah Code handling a repository change in its terminal interface](https://raw.githubusercontent.com/skundu42/noah-code/main/docs/assets/noah-in-action.svg)
 
 <p align="center"><sub>The real Textual interface, captured from a deterministic Noah Code session.</sub></p>
 
@@ -49,10 +51,10 @@ configuration or session history.
   preimages, concurrent-change detection, immediate diagnostics, and rollback.
 - **Visible execution.** Stream bounded shell output while commands run, keep servers and watchers
   as managed background jobs, and revisit activity details with `F2`.
-- **Recoverable work.** Review staged and unstaged changes in `/diff`, then undo or redo journaled
-  file edits even after restarting Noah.
-- **Persistent sessions.** Resume repository-scoped conversations, todos, model choices, and
-  automatically compacted context without losing the full underlying tool results.
+- **Crash-safe work.** Durable file intents roll back interrupted workspace-tool writes, Git
+  checkpoints protect shell-driven changes, and interrupted model runs resume after restart.
+- **Persistent sessions.** Resume repository-scoped conversations, steering, todos, model choices,
+  budgets, background-job logs, and compacted context without losing full tool results.
 - **Token-efficient by construction.** Lean tool-output bounds with disk-backed recall, condensed
   subagent results, cache-stable request prefixes (volatile status arrives as appended events),
   and pointer-eviction compaction — measured live with `/tokens`.
@@ -60,6 +62,9 @@ configuration or session history.
   **plan** mode, with ordered `allow`, `ask`, and `deny` permission rules.
 - **Extensible workflows.** Add slash commands, opt-in skills, MCP servers, or markdown subagents;
   attach `@files` and images when the task needs more context.
+- **Long-running by design.** Provider retries and configurable fallback models handle transient
+  failures, while workspace leases, bounded artifacts, durable process ownership, and `/health`
+  keep unattended sessions observable.
 
 Noah follows repository instructions from `AGENTS.md`, `CLAUDE.md`, and
 `.noah-code/instructions.md`.
@@ -116,6 +121,10 @@ Type `/` to search the full command and configuration reference. The most common
 | --- | --- |
 | `Enter` | Send the current prompt or accept a selected suggestion |
 | `Shift+Enter` | Insert a newline |
+| Drag, then `Cmd+C` / `Ctrl+Shift+C` | Select and copy TUI text |
+| `Ctrl+Shift+C` with no selection | Copy the latest Noah reply |
+| `Ctrl+G` | Open the searchable skills picker |
+| `Ctrl+]` | Return to live transcript output |
 | `Tab` | Switch between build and plan mode |
 | `F2` | Open execution activity |
 | `F3` | Open paginated conversation history |
@@ -123,13 +132,32 @@ Type `/` to search the full command and configuration reference. The most common
 | `/theme` | Choose Atom One Dark, Noah Ocean, Graphite, or High Contrast |
 | `/diff` | Review staged and unstaged changes |
 | `/undo` / `/redo` | Traverse the persistent edit journal |
+| `/checkpoints` | List rolling Git worktree checkpoints |
+| `/health` | Inspect durable run, job, inbox, event, and artifact state |
 | `/tokens` | Inspect tokens, cache usage, prefix stability, model wait, and tool output |
 | `/efficiency` | Switch between `fast`, `balanced`, and `deep` budgets |
 
 On wide terminals, the side rail prioritizes the active operation, Git branch and change counts,
 session, model usage, update state, and plan. Git status is refreshed in the background at turn
 boundaries, so the animated working state stays responsive. The main pane remains centered on the
-Noah mark until the first prompt, then becomes the conversation and execution timeline.
+large Noah wordmark until the first prompt, then becomes the conversation and execution timeline.
+
+## Sessions and crash recovery
+
+Start a new session with `noah .`, resume the latest repository session with
+`noah --continue .`, or reopen an exact session with `noah --session SESSION_ID .`. Inside the TUI,
+use `Ctrl+O`, `/sessions`, or `/continue`.
+
+Noah stores conversational history and host runtime state separately. If the process stops during
+an active run, the next launch restores pending steering, usage and budget counters, durable job
+logs, and the original request. Incomplete workspace-tool writes are rolled back, verified orphan
+process groups are cleaned up, and non-interactive runs continue automatically. A request that was
+waiting for user input remains paused for the next user message.
+
+By default, only one Noah process may own a checkout at a time. Use `/worktree create` when
+independent agents need to work concurrently. See
+[Reliability and long-running sessions](docs/reliability.md) for the recovery model, provider retry
+controls, quotas, and operational limits.
 
 ## Models and providers
 
@@ -158,7 +186,8 @@ noah --model openai/MODEL --reasoning-effort high .
 ```
 
 See the [provider configuration guide](docs/configuration.md#bring-your-own-api-provider) for
-gateway-specific setup.
+gateway-specific setup. Provider request deadlines, exponential retry, and ordered fallback models
+are configured under `[reliability.retries]`.
 
 ## Updates
 
@@ -175,6 +204,7 @@ noah update
 
 - [Interactive interface and sessions](docs/interactive-reference.md)
 - [Configuration, modes, permissions, and updates](docs/configuration.md)
+- [Reliability and long-running sessions](docs/reliability.md)
 - [Generated-code security](docs/security.md)
 - [Custom commands, skills, MCP, and tracing](docs/extensions.md)
 - [Development, CI, and releases](docs/development.md)

noah_code-0.2.4/docs/assets/noah-logo.svg

@@ -0,0 +1,15 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 760 230" role="img" aria-labelledby="title description">
+  <title id="title">Noah Code</title>
+  <desc id="description">The six-line Noah terminal wordmark used by the Noah Code TUI.</desc>
+  <rect width="760" height="230" rx="22" fill="#07151d"/>
+  <rect x="1" y="1" width="758" height="228" rx="21" fill="none" stroke="#173746" stroke-width="2"/>
+  <g fill="#e6edf3" font-family="ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, Liberation Mono, monospace" font-size="20" font-weight="700" text-anchor="middle">
+    <text x="380" y="43">███╗   ██╗ ██████╗  █████╗ ██╗  ██╗</text>
+    <text x="380" y="69">████╗  ██║██╔═══██╗██╔══██╗██║  ██║</text>
+    <text x="380" y="95">██╔██╗ ██║██║   ██║███████║███████║</text>
+    <text x="380" y="121">██║╚██╗██║██║   ██║██╔══██║██╔══██║</text>
+    <text x="380" y="147">██║ ╚████║╚██████╔╝██║  ██║██║  ██║</text>
+    <text x="380" y="173">╚═╝  ╚═══╝ ╚═════╝ ╚═╝  ╚═╝╚═╝  ╚═╝</text>
+  </g>
+  <text x="380" y="207" fill="#5bd1d7" font-family="ui-monospace, SFMono-Regular, Menlo, Monaco, Consolas, Liberation Mono, monospace" font-size="16" font-weight="700" letter-spacing="5" text-anchor="middle">NOAH / C O D E</text>
+</svg>

{noah_code-0.2.3 → noah_code-0.2.4}/docs/configuration.md

@@ -112,6 +112,8 @@ max_output_lines = 250
 max_search_results = 100
 max_file_results = 500
 tool_output_retention_hours = 24
+subagent_result_max_chars = 4000
+max_concurrent_subagents = 3
 
 [lsp]
 enabled = true
@@ -122,15 +124,51 @@ max_symbols = 300
 
 [processes]
 max_jobs = 8
-max_runtime_seconds = 3600
+max_runtime_seconds = 86400 # 24 hours
 max_buffer_chars = 64000
 stop_grace_seconds = 2
 
+[sampling]
+# Omit values to use provider defaults.
+# temperature = 0.2
+# top_p = 0.95
+# seed = 42
+
+[budget]
+# Persistent session-wide limits; omitted values are unlimited.
+# max_tokens = 500000
+# max_cost_usd = 25
+# max_seconds = 28800
+
+[checkpoints]
+enabled = true
+max_per_session = 50
+capture_before_mutation = true
+
+[reliability]
+auto_resume_interrupted_runs = true
+interaction_timeout_seconds = 86400
+artifact_max_bytes = 2000000000
+session_max_bytes = 5000000000
+max_runtime_events = 20000
+workspace_lease = true
+
+[reliability.retries]
+max_attempts = 5
+base_delay_seconds = 0.5
+max_delay_seconds = 20
+jitter_ratio = 0.2
+request_timeout_seconds = 180
+fallback_models = []
+
+[hooks]
+# pre_tool = [{ match = "ws_*", command = "./scripts/pre-tool-check" }]
+# post_tool = [{ match = "*", command = "./scripts/audit-tool" }]
+
 [ui]
 theme = "atom-one-dark" # atom-one-dark, noah-ocean, graphite, or high-contrast
 frontend = "tui"       # "tui" or "console"
 markdown = true
-stream_shell = true
 show_reasoning = false
 
 [summarization]
@@ -141,7 +179,7 @@ target_chars = 2500
 
 [tracing]
 enabled = true
-viewer = true
+# By default JSONL traces live inside each session directory.
 # jsonl_dir = "~/.local/share/noah-code/traces"
 
 [updates]
@@ -162,10 +200,12 @@ Supported environment overrides include:
 - `NOAH_CODE_AUTO_UPDATE`
 
 Repository-controlled configuration cannot weaken the host trust boundary. Project config is
-ignored for `auto_approve`, `efficiency`, `enabled_skills`, `lsp`, `mcp`, `permission_rules`,
-`processes`, `session_dir`, `tracing`, `updates`, and `unsafe_inprocess_code_execution`. Put those
-settings in trusted user config, the environment, or an explicit CLI flag. Language-server
-overrides are user-only because they launch local executables.
+ignored for `auto_approve`, `budget`, `efficiency`, `enabled_skills`, `hooks`, `lsp`, `mcp`,
+`permission_rules`, `processes`, `reliability`, `session_dir`, `tracing`, `updates`, and
+`unsafe_inprocess_code_execution`. Put those settings in trusted user config, the environment, or
+an explicit CLI flag. Language-server overrides and hooks are user-only because they launch local
+executables; reliability and budget settings are user-only so repository content cannot weaken
+host limits.
 
 A user-configured `permission_rules` array replaces the default rule array. Copy forward every
 default you still want before adding overrides. Hard secret, destructive-shell, and plan-mode
@@ -175,6 +215,10 @@ Inspect the resolved configuration from the CLI or inside an interactive session
 lists every nested path, while an optional path scopes the output. Values whose names look like
 credentials are redacted.
 
+Configuration is validated strictly: unknown keys (including typos such as `theem` under `[ui]`)
+and invalid values or TOML syntax fail every command with a one-line error naming the file and
+field, never a traceback. Fix the named key or remove it; `noah doctor` reports the same error.
+
 ```bash
 noah config show .
 ```
@@ -219,40 +263,67 @@ safety rail (default 40), not an efficiency-profile cap. Switch without restarti
 /efficiency deep
 ```
 
-Oversized results are not discarded. Noah writes the exact output to a private cache file for the
-configured retention period, returns a bounded head/tail preview, and gives the agent an output ID
-for focused line-range retrieval. A truncated file preview is never returned as an editable Match
-anchor.
+Oversized results are not discarded. During a durable session Noah writes the exact output to a
+private, content-addressed artifact store, returns a bounded head/tail preview, and gives the agent
+an output ID for focused line-range retrieval. Artifacts remain available when the session is
+resumed and count toward `reliability.artifact_max_bytes` and `session_max_bytes`. A truncated file
+preview is never returned as an editable Match anchor. `tool_output_retention_hours` applies to the
+fallback cache used when workspace tools are embedded without a durable session runtime.
 
 Set `lightweight_model` to route compaction to a faster or cheaper model. If it is omitted, that
 route follows live `/model` switches. Compaction starts at 35% of the active main model's context
 window by default, preserves the six newest events, and writes a coding checkpoint covering the
 objective, decisions, files, validation, blockers, and next steps.
 
+### Budgets, checkpoints, and reliability
+
+`[budget]` limits are cumulative across the parent agent, lightweight route, and custom-model
+subagents. Token, cost, and elapsed-time counters survive process restarts. When a configured cap
+is exceeded, the breach is sticky and later model calls fail before contacting the provider.
+
+Git checkpoints are enabled by default and use rolling retention: once the configured maximum is
+reached, Noah removes the oldest ref and continues capturing instead of silently stopping.
+`capture_before_mutation` protects shell-driven changes; workspace-tool edits additionally use
+durable pre-images and the persistent `/undo` journal.
+
+`[reliability.retries]` controls transient model-call retries and ordered fallback models. Noah does
+not retry authentication, invalid-request, content-policy, or context-window failures through this
+route. `[reliability]` also controls crash-run resumption, interaction timeouts, session and artifact
+quotas, event retention, and the exclusive checkout lease. See
+[Reliability and long-running sessions](reliability.md) for the recovery sequence and operational
+limits.
+
 ## Modes and permissions
 
 | Mode | Behavior |
 |------|----------|
-| `build` | Reads are allowed; edits and shell commands follow permission rules and ask by default |
+| `build` | Routine workspace reads, edits, inspection, web research, skills, and delegation proceed automatically; higher-risk operations ask |
 | `plan` | Reads are allowed; file edits and mutating shell commands are denied. `self.plan.write` may pin `.noah-code/plan.md` |
 
 Switch modes with `--mode`, `/mode build`, or `/mode plan`. The agent can propose a switch with
-`self.plan.enter()` / `self.plan.exit_to_build()` after writing a plan. The active mode is stored with the
-session.
+`self.plan.enter()` / `self.plan.exit_to_build()` after writing a plan; the transition always asks
+for confirmation and is never auto-approved, so in `noah run --auto` it is refused and the session
+stays in plan mode. The active mode is stored with the session.
 
 Permission rules are evaluated in order, and the last matching rule wins. The default policy:
 
 - Allows ordinary reads.
-- Denies likely secrets, including `.env` variants, private keys, `.git` internals, and session
-  databases. `.env.example` remains readable.
-- Asks before workspace edits, shell commands, web fetches, web searches, and subagents.
+- Denies likely secrets, including `.env` variants, private keys, `.git` internals, credential
+  stores (`.npmrc`, `.pypirc`, `.netrc`, `.pgpass`, `.kube/config`, `.docker/config.json`,
+  `.aws/credentials`), Java/JCEKS key stores, and session databases. `.env.example` remains
+  readable. The same denials apply to Git object syntax (`git show HEAD:.env`) and to patch-output
+  Git commands with no path scope (`git log -p`, bare `git show`), which ask instead of being
+  auto-approved.
+- Allows workspace edits, read-only in-workspace shell inspection, web reads, skills, and
+  delegated tasks without interrupting the turn.
+- Asks before arbitrary shell execution, external paths, MCP access, and remote mutations.
 - Allows the question tool so the agent can pause for a structured choice.
 - Denies `git push`, `git clean`, `git reset --hard`, and mutating `gh pr`
   (`create`, `checkout`, `merge`, `close`, `ready`, `review`). Push and PR
   mutations go through `/pr` / `self.github` instead.
 - Allows listing and viewing pull requests; asks before create, push, checkout,
   or comment.
-- Keeps file tools inside the active workspace and asks before skill or MCP access.
+- Keeps file tools inside the active workspace and asks before MCP access.
 - Denies plan-mode mutations regardless of broader allow rules. Plan mode may still run
   read-only subagents. Its shell allowlist accepts only literal, unqualified read-only programs;
   test collection and interpreter execution are not treated as read-only because they can load
@@ -260,10 +331,12 @@ Permission rules are evaluated in order, and the last matching rule wins. The de
 
 `--auto` changes routine ask decisions to allow but never overrides an explicit deny.
 Elevated-risk commands such as file removal, downloads, and package installation still require
-explicit approval. Compound shell commands and mutating or unrecognized Git commands cannot be
-silently auto-approved. Interpreters, eval/source commands, and indirect execution wrappers, plus
-arguments hidden behind variable, command, ANSI-C, or brace expansion, are denied under `--auto`;
-run without `--auto` when one of these commands needs explicit approval.
+explicit approval: interactively they ask, and in non-interactive `noah run --auto` they are
+rejected outright instead of being silently approved. Compound shell commands and mutating or
+unrecognized Git commands cannot be silently auto-approved. Interpreters, eval/source commands, and
+indirect execution wrappers, plus arguments hidden behind variable, command, ANSI-C, or brace
+expansion, are denied under `--auto`; run without `--auto` when one of these commands needs
+explicit approval.
 
 ## Installation and updates
 

{noah_code-0.2.3 → noah_code-0.2.4}/docs/development.md

@@ -12,7 +12,8 @@ uv run pytest tests
 uv build
 ```
 
-The default test suite is hermetic and does not require network access or provider keys. A single
+The default test suite is hermetic and does not require network access or provider keys; pytest-socket
+blocks network sockets outright (only local AF_UNIX sockets, e.g. asyncio's self-pipe, stay open). A single
 opt-in test performs a live HTTP fetch; run it explicitly on a connected machine:
 
 ```bash
@@ -25,7 +26,7 @@ Measure coverage (a 70% total gate runs in CI):
 uv run pytest tests --cov=noah_code --cov-report=term-missing
 ```
 
-Optional git hooks mirror the lint gate:
+Optional git hooks mirror the lint, type-check, and lockfile gates:
 
 ```bash
 uvx pre-commit install
@@ -37,16 +38,19 @@ tool output, and estimated cost.
 ## CI
 
 GitHub Actions runs the complete test suite on Python 3.12 and 3.13, plus platform smoke tests on
-Linux and macOS for arm64 and x86_64. Every pull request and push to `main` must pass lint, static
-type checking (mypy), tests with a 70% coverage floor, lockfile validation, and a package build.
-Concurrent runs for the same reference are cancelled automatically.
+Linux and macOS for arm64 and x86_64. Every pull request, push to `main`, and `v*` tag must pass lint,
+static type checking (mypy), tests with a 70% coverage floor (measured on Ubuntu; the macOS smoke jobs
+also collect coverage, without a gate, to include mac-only paths), lockfile validation, and a package
+build. A separate job runs the network-dependent integration tests and is allowed to fail, since live
+network calls can flake for reasons unrelated to the code. Concurrent runs for the same reference are
+cancelled automatically.
 
 ## Releases
 
 A `v*` tag starts the release pipeline, which:
 
 1. Verifies that the tag, `pyproject.toml`, and package versions match.
-2. Reruns tests and builds the wheel and source distribution.
+2. Reruns lint, type checks, and tests, then builds the wheel and source distribution.
 3. Validates distribution metadata and generates SHA-256 checksums.
 4. Publishes to PyPI using short-lived OIDC credentials.
 5. Creates provenance attestations and a GitHub release with the artifacts.
@@ -61,8 +65,9 @@ when it exists and otherwise falls back to generated notes. Commit and push the
 a matching annotated tag. For example:
 
 ```bash
-git tag -a v0.2.0 -m "Noah Code v0.2.0"
-git push origin v0.2.0
+VERSION=v0.2.3
+git tag -a "$VERSION" -m "Noah Code $VERSION"
+git push origin "$VERSION"
 ```
 
 Once PyPI publishes the version, existing uv-tool installations discover it through the built-in