ce-tool 0.1.0__tar.gz

ce_tool-0.1.0/.claude/settings.json

@@ -0,0 +1,14 @@
+{
+  "hooks": {
+    "PreCompact": [
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "\"C:\\\\Python313\\\\python.exe\" -c \"from pathlib import Path; files = [('=== CRITICAL CONTEXT ===', Path('CLAUDE.md')), ('=== CURRENT TASKS ===', Path('TASKS.md'))]; [(print(title), print(path.read_text(encoding='utf-8', errors='replace'))) for title, path in files if path.exists()]\""
+          }
+        ]
+      }
+    ]
+  }
+}

ce_tool-0.1.0/.claudeignore

@@ -0,0 +1,28 @@
+# Auto-generated by claude-efficient
+
+.git/
+*.log
+.DS_Store
+Thumbs.db
+*.png
+*.jpg
+*.jpeg
+*.gif
+*.pdf
+*.bin
+*.safetensors
+*.pt
+*.onnx
+
+# python
+**/__pycache__/**
+**/*.pyc
+**/*.pyo
+*.egg-info/
+dist/
+build/
+.venv/
+venv/
+.ruff_cache/
+.mypy_cache/
+.pytest_cache/

ce_tool-0.1.0/.geminiignore

@@ -0,0 +1,28 @@
+# Auto-generated by claude-efficient
+
+.git/
+*.log
+.DS_Store
+Thumbs.db
+*.png
+*.jpg
+*.jpeg
+*.gif
+*.pdf
+*.bin
+*.safetensors
+*.pt
+*.onnx
+
+# python
+**/__pycache__/**
+**/*.pyc
+**/*.pyo
+*.egg-info/
+dist/
+build/
+.venv/
+venv/
+.ruff_cache/
+.mypy_cache/
+.pytest_cache/

ce_tool-0.1.0/.github/workflows/ci.yml

@@ -0,0 +1,15 @@
+name: CI
+
+on: [push, pull_request]
+
+jobs:
+  test:
+    runs-on: ubuntu-22.04
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+      - run: pip install -e ".[dev]"
+      - run: ruff check .
+      - run: pytest tests/ -x

ce_tool-0.1.0/AGENTS.md

@@ -0,0 +1,56 @@
+# Codex-efficient
+
+CLI tool that wraps Codex sessions to eliminate token waste. `ce` command.
+
+## Architecture
+```
+cli/            # main.py (Click group), init.py, session.py, audit.py, commands.py
+generators/     # claude_md.py, claudeignore.py, backends.py
+session/        # model_router.py, compact_manager.py, mcp_config.py, subagent_planner.py
+analysis/       # cache_health.py, waste_detector.py
+prompt/         # optimizer.py
+config/         # defaults.toml, schema.py
+tests/          # mirrors src/ structure
+```
+
+## Key files
+- `config/defaults.toml` — all thresholds; always_keep guards Codex-mem hooks
+- `generators/backends.py` — GeminiBackend, OllamaBackend, LLMBackend, ClaudeBackend + detect_backend()
+- `generators/claude_md.py` — ClaudeMdGenerator: scan, generate, write, @import tree, PreCompact hook
+- `cli/main.py` — Click group; all commands registered here
+- `session/model_router.py` — model selected ONCE at session start; inject_model_flag()
+- `session/mcp_config.py` — advises on ENABLE_EXPERIMENTAL_MCP_CLI; never removes servers
+- `analysis/cache_health.py` — detects prompt cache prefix violations before they happen
+
+## Specs (load the relevant one only — do NOT read all at once)
+specs/SPEC_01_04_FIXES.md     # architecture corrections for already-built code
+specs/SPEC_05_06_revised.md   # cache_health.py + updated ce init
+specs/SPEC_07_11_revised.md   # session layer: router, compact, mcp_config, optimizer, ce run
+specs/SPEC_12_16_revised.md   # audit, subagents, commands, packaging
+
+## Architecture constraints (non-negotiable)
+- **Model routing is session-start-only.** Switching models mid-session invalidates the
+  prompt cache prefix — costing more than running Opus start-to-finish.
+- **MCP servers are never removed mid-session.** Use ENABLE_EXPERIMENTAL_MCP_CLI=true
+  for on-demand schema loading. always_keep = ["claude_mem", "memory", "filesystem"].
+- **claude_mem hooks are not user-invoked tools.** PostToolUse, SessionEnd fire passively.
+  Disabling the server kills the capture pipeline; that entire session produces zero memory.
+- **Compact threshold: 45%.** Default action is /clear + fresh session, not /compact.
+  Two compactions in one session = task was scoped too large.
+
+## Run commands
+```bash
+pip install -e ".[dev]"
+ce --help
+pytest tests/ -x
+ruff check .
+```
+
+## Current phase
+See TASKS.md
+
+## Output format
+- Code only. No preamble. No "let me first check" narration.
+- One file per response unless explicitly asked for multiple.
+- Bash verification at the end only, never mid-write.
+- Imports at top, no inline explanations inside code.

ce_tool-0.1.0/CLAUDE.md

@@ -0,0 +1,26 @@
+```json
+{
+  "project_digest_root": {
+    "old_mds": {},
+    "src": {
+      "claude_efficient": {
+        "cli": {},
+        "generators": {},
+        "session": {}
+      }
+    },
+    "tests": {},
+    "AGENTS.md": {},
+    "README.md": {},
+    "Screenshot 2026-03-20 112524.png": {},
+    "pyproject.toml": {},
+    "settings.json": {}
+  }
+}
+```
+## Subdirectory context (auto-loaded by Claude Code)
+
+@src/claude_efficient/cli/CLAUDE.md
+@src/claude_efficient/generators/CLAUDE.md
+@src/claude_efficient/session/CLAUDE.md
+@tests/CLAUDE.md

ce_tool-0.1.0/GEMINI.md

@@ -0,0 +1,26 @@
+```json
+{
+  "project_digest_root": {
+    "old_mds": {},
+    "src": {
+      "claude_efficient": {
+        "cli": {},
+        "generators": {},
+        "session": {}
+      }
+    },
+    "tests": {},
+    "AGENTS.md": {},
+    "README.md": {},
+    "Screenshot 2026-03-20 112524.png": {},
+    "pyproject.toml": {},
+    "settings.json": {}
+  }
+}
+```
+## Subdirectory context (auto-loaded by Claude Code)
+
+@src/claude_efficient/cli/GEMINI.md
+@src/claude_efficient/generators/GEMINI.md
+@src/claude_efficient/session/GEMINI.md
+@tests/GEMINI.md

ce_tool-0.1.0/PKG-INFO

@@ -0,0 +1,81 @@
+Metadata-Version: 2.4
+Name: ce-tool
+Version: 0.1.0
+Summary: Token optimization wrapper for Claude Code sessions
+Project-URL: Homepage, https://github.com/thedotmack/claude-efficient
+Project-URL: Issues, https://github.com/thedotmack/claude-efficient/issues
+Requires-Python: >=3.11
+Requires-Dist: click>=8.1
+Requires-Dist: requests>=2.31
+Requires-Dist: rich>=13.0
+Requires-Dist: tomllib>=1.0; python_version < '3.11'
+Provides-Extra: dev
+Requires-Dist: build; extra == 'dev'
+Requires-Dist: pytest-mock>=3.0; extra == 'dev'
+Requires-Dist: pytest>=8.0; extra == 'dev'
+Requires-Dist: ruff>=0.4; extra == 'dev'
+Description-Content-Type: text/markdown
+
+<div align="center">
+  <br />
+  <h1>Claude Efficient</h1>
+  <p>
+    <strong>Stop burning Claude Code tokens.</strong>
+  </p>
+  <p>
+    <code>ce</code> wraps every session with automatic optimizations to save you money and improve context quality.
+  </p>
+</div>
+
+---
+
+## 🚀 Installation
+
+`claude-efficient` is a Python package. Ensure you have Python 3.11+ installed.
+
+```bash
+pip install claude-efficient
+```
+
+This will install the `ce` command-line tool.
+
+## ✨ Quick Start
+
+1.  **Initialize your project:**
+    ```bash
+    ce init
+    ```
+    This command sets up your project with a `CLAUDE.md`, `.claudeignore`, and other necessary configurations for optimal performance.
+
+2.  **Run a task:**
+    ```bash
+    ce run "Your task description here. For example, build a FastAPI endpoint for user auth."
+    ```
+    This is the main command you'll use. It automatically optimizes your request, manages session context, and routes it appropriately.
+
+3.  **Check your savings:**
+    ```bash
+    ce gains
+    ```
+    This command displays a detailed dashboard of your token savings across different operations.
+
+## 📋 Commands
+
+`claude-efficient` provides a suite of commands to manage your workflow and optimize token usage.
+
+| Command         | Description                                                              |
+| --------------- | ------------------------------------------------------------------------ |
+| `ce init`       | Initializes a new or existing project for use with `claude-efficient`.   |
+| `ce run [TASK]` | Runs a new coding session with the given task, applying optimizations.   |
+| `ce gains`      | Displays the token savings dashboard.                                    |
+| `ce status`     | Shows a health dashboard for the current project configuration.          |
+| `ce audit [LOG]`| Audits a session log to detect inefficiencies and suggest improvements.  |
+| `ce telemetry`  | Shows before/after token savings from recorded sessions.                 |
+| `ce mem-search` | Searches cross-session memory for relevant context from past tasks.      |
+| `ce scope-check`| Estimates the token requirements for a task before running a session.    |
+| `ce helpers`    | Provides assistance with setting up helpers and integrations.            |
+
+---
+<div align="center">
+  <p>Made with ❤️ for efficient coding sessions.</p>
+</div>

ce_tool-0.1.0/README.md

@@ -0,0 +1,63 @@
+<div align="center">
+  <br />
+  <h1>Claude Efficient</h1>
+  <p>
+    <strong>Stop burning Claude Code tokens.</strong>
+  </p>
+  <p>
+    <code>ce</code> wraps every session with automatic optimizations to save you money and improve context quality.
+  </p>
+</div>
+
+---
+
+## 🚀 Installation
+
+`claude-efficient` is a Python package. Ensure you have Python 3.11+ installed.
+
+```bash
+pip install claude-efficient
+```
+
+This will install the `ce` command-line tool.
+
+## ✨ Quick Start
+
+1.  **Initialize your project:**
+    ```bash
+    ce init
+    ```
+    This command sets up your project with a `CLAUDE.md`, `.claudeignore`, and other necessary configurations for optimal performance.
+
+2.  **Run a task:**
+    ```bash
+    ce run "Your task description here. For example, build a FastAPI endpoint for user auth."
+    ```
+    This is the main command you'll use. It automatically optimizes your request, manages session context, and routes it appropriately.
+
+3.  **Check your savings:**
+    ```bash
+    ce gains
+    ```
+    This command displays a detailed dashboard of your token savings across different operations.
+
+## 📋 Commands
+
+`claude-efficient` provides a suite of commands to manage your workflow and optimize token usage.
+
+| Command         | Description                                                              |
+| --------------- | ------------------------------------------------------------------------ |
+| `ce init`       | Initializes a new or existing project for use with `claude-efficient`.   |
+| `ce run [TASK]` | Runs a new coding session with the given task, applying optimizations.   |
+| `ce gains`      | Displays the token savings dashboard.                                    |
+| `ce status`     | Shows a health dashboard for the current project configuration.          |
+| `ce audit [LOG]`| Audits a session log to detect inefficiencies and suggest improvements.  |
+| `ce telemetry`  | Shows before/after token savings from recorded sessions.                 |
+| `ce mem-search` | Searches cross-session memory for relevant context from past tasks.      |
+| `ce scope-check`| Estimates the token requirements for a task before running a session.    |
+| `ce helpers`    | Provides assistance with setting up helpers and integrations.            |
+
+---
+<div align="center">
+  <p>Made with ❤️ for efficient coding sessions.</p>
+</div>

ce_tool-0.1.0/old_mds/CLAUDE.md

@@ -0,0 +1,56 @@
+# claude-efficient
+
+CLI tool that wraps Claude Code sessions to eliminate token waste. `ce` command.
+
+## Architecture
+```
+cli/            # main.py (Click group), init.py, session.py, audit.py, commands.py
+generators/     # claude_md.py, claudeignore.py, backends.py
+session/        # model_router.py, compact_manager.py, mcp_config.py, subagent_planner.py
+analysis/       # cache_health.py, waste_detector.py
+prompt/         # optimizer.py
+config/         # defaults.toml, schema.py
+tests/          # mirrors src/ structure
+```
+
+## Key files
+- `config/defaults.toml` — all thresholds; always_keep guards claude-mem hooks
+- `generators/backends.py` — GeminiBackend, OllamaBackend, LLMBackend, ClaudeBackend + detect_backend()
+- `generators/claude_md.py` — ClaudeMdGenerator: scan, generate, write, @import tree, PreCompact hook
+- `cli/main.py` — Click group; all commands registered here
+- `session/model_router.py` — model selected ONCE at session start; inject_model_flag()
+- `session/mcp_config.py` — advises on ENABLE_EXPERIMENTAL_MCP_CLI; never removes servers
+- `analysis/cache_health.py` — detects prompt cache prefix violations before they happen
+
+## Specs (load the relevant one only — do NOT read all at once)
+specs/SPEC_01_04_FIXES.md     # architecture corrections for already-built code
+specs/SPEC_05_06_revised.md   # cache_health.py + updated ce init
+specs/SPEC_07_11_revised.md   # session layer: router, compact, mcp_config, optimizer, ce run
+specs/SPEC_12_16_revised.md   # audit, subagents, commands, packaging
+
+## Architecture constraints (non-negotiable)
+- **Model routing is session-start-only.** Switching models mid-session invalidates the
+  prompt cache prefix — costing more than running Opus start-to-finish.
+- **MCP servers are never removed mid-session.** Use ENABLE_EXPERIMENTAL_MCP_CLI=true
+  for on-demand schema loading. always_keep = ["claude_mem", "memory", "filesystem"].
+- **claude_mem hooks are not user-invoked tools.** PostToolUse, SessionEnd fire passively.
+  Disabling the server kills the capture pipeline; that entire session produces zero memory.
+- **Compact threshold: 45%.** Default action is /clear + fresh session, not /compact.
+  Two compactions in one session = task was scoped too large.
+
+## Run commands
+```bash
+pip install -e ".[dev]"
+ce --help
+pytest tests/ -x
+ruff check .
+```
+
+## Current phase
+See TASKS.md
+
+## Output format
+- Code only. No preamble. No "let me first check" narration.
+- One file per response unless explicitly asked for multiple.
+- Bash verification at the end only, never mid-write.
+- Imports at top, no inline explanations inside code.