patchpal 0.21.2__tar.gz → 0.21.4__tar.gz

{patchpal-0.21.2/patchpal.egg-info → patchpal-0.21.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: patchpal
-Version: 0.21.2
+Version: 0.21.4
 Summary: An agentic coding and automation assistant, supporting both local and cloud LLMs
 Author: PatchPal Contributors
 License-Expression: Apache-2.0
@@ -79,6 +79,32 @@ $ patchpal              # start
 
 > Platform support: Linux, macOS, and Windows are all supported
 
+**Alternative: Run with Docker/Podman (no installation required)**
+
+```bash
+# Using pre-built image with patchpal installed (default model)
+docker run -it --rm \
+  -v $(pwd):/workspace \
+  -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal
+
+# Or with Podman
+podman run -it --rm \
+  -v $(pwd):/workspace \
+  -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal
+
+# Specify a different model with --model
+docker run -it --rm \
+  -v $(pwd):/workspace \
+  -e OPENAI_API_KEY=$OPENAI_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal --model openai/gpt-5-mini
+```
+
+
 ## Setup
 0. **Install**: `pip install patchpal`
 1. **Get an API key or a Local LLM Engine**:

{patchpal-0.21.2 → patchpal-0.21.4}/README.md

@@ -30,6 +30,32 @@ $ patchpal              # start
 
 > Platform support: Linux, macOS, and Windows are all supported
 
+**Alternative: Run with Docker/Podman (no installation required)**
+
+```bash
+# Using pre-built image with patchpal installed (default model)
+docker run -it --rm \
+  -v $(pwd):/workspace \
+  -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal
+
+# Or with Podman
+podman run -it --rm \
+  -v $(pwd):/workspace \
+  -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal
+
+# Specify a different model with --model
+docker run -it --rm \
+  -v $(pwd):/workspace \
+  -e OPENAI_API_KEY=$OPENAI_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal --model openai/gpt-5-mini
+```
+
+
 ## Setup
 0. **Install**: `pip install patchpal`
 1. **Get an API key or a Local LLM Engine**:

{patchpal-0.21.2 → patchpal-0.21.4}/patchpal/__init__.py

@@ -1,6 +1,6 @@
 """PatchPal - An open-source Claude Code clone implemented purely in Python."""
 
-__version__ = "0.21.2"
+__version__ = "0.21.4"
 
 from patchpal.agent import create_agent, create_react_agent
 from patchpal.cli.autopilot import autopilot_loop

{patchpal-0.21.2 → patchpal-0.21.4}/patchpal/cli/sandbox.py

@@ -247,7 +247,14 @@ def build_container_args(sandbox_args, patchpal_args):
     import shlex
 
     quoted_args = " ".join(shlex.quote(arg) for arg in patchpal_cmd_args)
-    shell_cmd = f"pip install -q patchpal && {patchpal_cmd} {quoted_args}"
+
+    # If using pre-built patchpal-sandbox image, skip pip install
+    # Otherwise, install patchpal first (fallback for custom images)
+    if "patchpal-sandbox" in sandbox_args.image:
+        shell_cmd = f"{patchpal_cmd} {quoted_args}"
+    else:
+        shell_cmd = f"pip install -q patchpal && {patchpal_cmd} {quoted_args}"
+
     container_args.extend(["bash", "-c", shell_cmd])
 
     return container_args, runtime
@@ -269,7 +276,7 @@ DESCRIPTION:
     - Network access (required for cloud LLMs like OpenAI, Anthropic)
     - No resource limits by default (uses Docker/Podman defaults)
     - Current directory mounted as /workspace
-    - Auto-installs patchpal in container
+    - Pre-built image with patchpal installed (fast startup)
     - Auto-mounts ~/.patchpal for custom tools, config, and memory
     - Custom tools work automatically (from ~/.patchpal/tools/ and <repo>/.patchpal/tools/)
     - Auto-sets OLLAMA_CONTEXT_LENGTH for Ollama models:
@@ -279,7 +286,8 @@ DESCRIPTION:
     Recommended for autopilot mode and high-risk operations.
 
 SCRIPT OPTIONS:
-    --image IMAGE       Container image to use (default: python:3.11-slim)
+    --image IMAGE       Container image to use (default: ghcr.io/amaiya/patchpal-sandbox:latest)
+                        Use python:3.11-slim to get latest patchpal from PyPI (slower startup)
     --no-network        Disable network access (incompatible with cloud LLMs)
     --host-network      Use host network (for local Ollama/vLLM servers on localhost)
     --memory LIMIT      Memory limit (e.g., 2g, 4g) - optional, no limit by default
@@ -314,8 +322,9 @@ ENVIRONMENT VARIABLES:
         # Custom endpoints (optional)
         OPENAI_BASE_URL=https://your-proxy.com/v1
 
-        # Disable permissions for autopilot
-        PATCHPAL_REQUIRE_PERMISSION=false
+        # Optional: Disable permissions for interactive mode
+        # (Not needed for autopilot - it disables permissions automatically)
+        # PATCHPAL_REQUIRE_PERMISSION=false
 
     Note: SSL_CERT_FILE and REQUESTS_CA_BUNDLE paths are auto-mounted if they exist.
 
@@ -329,10 +338,15 @@ SECURITY:
     - No resource limits by default (trust Docker/Podman and OS limits)
     - Clean environment on each run (--rm flag)
     - Workspace files visible but container has limited privileges
+    - Permissions behavior:
+      * Interactive mode (default): Permissions ENABLED (prompts before operations)
+      * Autopilot mode: Permissions DISABLED automatically (autonomous operation)
 
 NOTES:
-    - First run downloads the Python image (~130MB)
-    - Each run reinstalls patchpal (takes ~10s)
+    - Default image (ghcr.io/amaiya/patchpal-sandbox:latest) has patchpal pre-installed for fast startup
+    - First run downloads the image (~150MB, one-time)
+    - Subsequent runs start instantly (no pip install needed)
+    - For latest patchpal from PyPI use: --image python:3.11-slim (slower, ~10-30s pip install)
     - Current directory mounted at /workspace (read-write)
     - ~/.patchpal mounted at /root/.patchpal if it exists
     - SSL certificates auto-mounted from /etc/ssl/certs if present
@@ -356,17 +370,18 @@ CORPORATE NETWORKS (Linux/WSL):
     You may need to add certificates to Docker Desktop's trusted CAs.
 
 EXAMPLES:
-    # Cloud LLM: Load API keys from .env file
-    patchpal-sandbox --env-file .env -- --model openai/gpt-5.2-codex
-    patchpal-sandbox --env-file ~/.config/patchpal/.env -- --model anthropic/claude-sonnet-4-5
 
-    #Cloud LLM: Interactive mode with cloud LLM (network enabled by default)
+    # Interactive mode - permissions ENABLED (prompts before operations)
     patchpal-sandbox -- --model anthropic/claude-sonnet-4-5
 
-    # AutoPilot Mode: non-interactive, permissions automatically disabled
-    patchpal-sandbox --env-file .env -- autopilot --model openai/gpt-5.2-codex --prompt "Add error handling to auth.py" --completion-promise "COMPLETE"
+    # Interactive mode - load API keys from .env file
+    patchpal-sandbox --env-file .env -- --model openai/gpt-5.2-codex
+    patchpal-sandbox --env-file ~/.config/patchpal/.env -- --model anthropic/claude-sonnet-4-5
+
+    # AutoPilot mode - permissions automatically DISABLED
+    patchpal-sandbox  -- autopilot --model openai/gpt-5-mini --prompt "Add error handling to auth.py" --completion-promise "COMPLETE"
 
-    # AutoPilot Mode: read file containing prompt
+    # AutoPilot mode - read file containing prompt and and .env file
     patchpal-sandbox --env-file .env -- autopilot --model openai/gpt-5.2-codex --prompt-file task.md --completion-promise "DONE"
 
     # Ollama: Linux/WSL requires host network to reach Ollama on localhost
@@ -411,7 +426,7 @@ def main():
 
     # Parse sandbox arguments
     parser = argparse.ArgumentParser(add_help=False)  # We handle help manually
-    parser.add_argument("--image", default="python:3.11-slim")
+    parser.add_argument("--image", default="ghcr.io/amaiya/patchpal-sandbox:latest")
     parser.add_argument("--network", default="bridge")
     parser.add_argument("--no-network", action="store_true")
     parser.add_argument("--host-network", action="store_true")

{patchpal-0.21.2 → patchpal-0.21.4/patchpal.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: patchpal
-Version: 0.21.2
+Version: 0.21.4
 Summary: An agentic coding and automation assistant, supporting both local and cloud LLMs
 Author: PatchPal Contributors
 License-Expression: Apache-2.0
@@ -79,6 +79,32 @@ $ patchpal              # start
 
 > Platform support: Linux, macOS, and Windows are all supported
 
+**Alternative: Run with Docker/Podman (no installation required)**
+
+```bash
+# Using pre-built image with patchpal installed (default model)
+docker run -it --rm \
+  -v $(pwd):/workspace \
+  -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal
+
+# Or with Podman
+podman run -it --rm \
+  -v $(pwd):/workspace \
+  -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal
+
+# Specify a different model with --model
+docker run -it --rm \
+  -v $(pwd):/workspace \
+  -e OPENAI_API_KEY=$OPENAI_API_KEY \
+  ghcr.io/amaiya/patchpal-sandbox:latest \
+  patchpal --model openai/gpt-5-mini
+```
+
+
 ## Setup
 0. **Install**: `pip install patchpal`
 1. **Get an API key or a Local LLM Engine**: