Convert gitgram/ and aitimemachine/ from embedded git repos (which showed as opaque submodule gitlinks on GitHub) into fully-tracked first-party content. AITM now merges its 7 vendored repos as first-party engines: - memory: beads (Go task DAG) + context7 (docs SDK), unified into saved_contexts - git: vendored C git (bundled binary) + gitnexus (in-process code intelligence) - agentic: a2a (@a2a-js/sdk) + pal (AITM-managed Python subprocess) Includes the full session's product build: Gitgram web/mobile/backend features, AITM website/CLI/MCP, the unified memory layer, engine build tooling (scripts/build-engines.sh, Dockerfile, CI), docs (README, ENGINES.md, ARCHITECTURE.md), and extensive test suites. Build artifacts, node_modules, external/, and gemini-cli are gitignored. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
89 lines
2.3 KiB
Markdown
89 lines
2.3 KiB
Markdown
# Python Agent Example
|
|
|
|
A simple Python script demonstrating how an AI agent can use bd to manage tasks.
|
|
|
|
## Features
|
|
|
|
- Finds ready work using `bd ready --json`
|
|
- Claims tasks by updating status
|
|
- Simulates discovering new issues during work
|
|
- Links discovered issues with `discovered-from` dependency
|
|
- Completes tasks and moves to the next one
|
|
|
|
## Prerequisites
|
|
|
|
- Python 3.7+
|
|
- bd installed: `curl -fsSL https://raw.githubusercontent.com/gastownhall/beads/main/scripts/install.sh | bash`
|
|
- A beads database initialized: `bd init`
|
|
|
|
## Usage
|
|
|
|
```bash
|
|
# Make the script executable
|
|
chmod +x agent.py
|
|
|
|
# Run the agent
|
|
./agent.py
|
|
```
|
|
|
|
## What It Does
|
|
|
|
1. Queries for ready work (no blocking dependencies)
|
|
2. Claims the highest priority task
|
|
3. "Works" on the task (simulated)
|
|
4. If the task involves implementation, discovers a testing task
|
|
5. Creates the new testing task and links it with `discovered-from`
|
|
6. Completes the original task
|
|
7. Repeats until no ready work remains
|
|
|
|
## Example Output
|
|
|
|
```
|
|
๐ Beads Agent starting...
|
|
|
|
============================================================
|
|
Iteration 1/10
|
|
============================================================
|
|
|
|
๐ Claiming task: bd-1
|
|
๐ค Working on: Implement user authentication (bd-1)
|
|
Priority: 1, Type: feature
|
|
|
|
๐ก Discovered: Missing test coverage for this feature
|
|
โจ Creating issue: Add tests for Implement user authentication
|
|
๐ Linking bd-2 โ discovered-from โ bd-1
|
|
โ
Completing task: bd-1 - Implemented successfully
|
|
|
|
๐ New work discovered and linked. Running another cycle...
|
|
```
|
|
|
|
## Integration with Real Agents
|
|
|
|
To integrate with a real LLM-based agent:
|
|
|
|
1. Replace `simulate_work()` with actual LLM calls
|
|
2. Parse the LLM's response for discovered issues/bugs
|
|
3. Use the issue ID to track context across conversations
|
|
4. Export/import JSONL to share state across agent sessions
|
|
|
|
## Advanced Usage
|
|
|
|
```python
|
|
# Create an agent with custom behavior
|
|
agent = BeadsAgent()
|
|
|
|
# Find specific types of work
|
|
ready = agent.run_bd("ready", "--priority", "1", "--assignee", "bot")
|
|
|
|
# Create issues with labels
|
|
agent.run_bd("create", "New task", "-l", "urgent,backend")
|
|
|
|
# Query dependency tree
|
|
tree = agent.run_bd("dep", "tree", "bd-1")
|
|
```
|
|
|
|
## See Also
|
|
|
|
- [../bash-agent/](../bash-agent/) - Bash version of this example
|
|
- [../claude-desktop-mcp/](../claude-desktop-mcp/) - MCP server for Claude Desktop
|