self_improving_coding_agent
A coding agent framework, that works on its own codebase.
File Explorer
Download Latest Version (.zip)- __init__.py
- archive_explorer.py
- coder.py
- main_orchestrator.py
- problem_solver.py
- reasoner.py
- review_committee_member.py
- __init__.py
- agent_calling.py
- assistant_base_agent.py
- base_agent.py
- __init__.py
- aime.py
- aiq_benchmark.py
- aiq_project_benchmarks.py
- arc_agi.py
- base.py
- drop.py
- file_editing.py
- gpqa.py
- gsm8k.py
- gsm_ic.py
- humaneval.py
- livecodebench.py
- math.py
- refute.py
- swebench_verified.py
- symbol_location.py
- __init__.py
- digraph.py
- manager.py
- reporting.py
- __init__.py
- event_bus.py
- event_bus_utils.py
- __init__.py
- anthropic.py
- base_provider.py
- deepseek.py
- fireworks.py
- google.py
- google_oai.py
- google_rest.py
- openai.py
- vertex.py
- __init__.py
- api.py
- base.py
- metering.py
- graph_visualisation.py
- overseer.py
- __init__.py
- json_parsing.py
- representation.py
- xml_dumps.py
- xml_parsing.py
- __init__.py
- overwrite_file.py
- utils.py
- __init__.py
- coding.py
- meta_improvement.py
- sequential.py
- sequential_subagents.py
- __init__.py
- answer_submission.py
- archive_tools.py
- base_agent_tools.py
- base_tool.py
- calculator.py
- committee_design.py
- directory_tools.py
- execute_command.py
- file_tools.py
- ripgrep_tool.py
- __init__.py
- agent_types.py
- common.py
- event_types.py
- llm_types.py
- tool_types.py
- __init__.py
- archive_analysis.py
- documentation.py
- file_views.py
- metrics.py
- parsing.py
- stop_tokens.py
- event-stream.js
- execution-tree.js
- metrics-display.js
- event-utils.js
- formatters.js
- core.js
- store.js
- styles.css
- visualizer.js
- index.html
- __init__.py
- server.py
- __init__.py
- config.py
- test_agent_calling.py
- disabled_test_livecode_benchmark.py
- test_benchmark_base.py
- test_file_editing.py
- test_gsm8k_benchmark.py
- test_refute_benchmark.py
- test_event_bus.py
- test_sequential.py
- test_base_tool.py
- test_calculator.py
- test_execute_command.py
- test_archive_analysis.py
- test_parsing.py
- test_example.py
- .gitignore
- __main__.py
- agent.py
- agent_change_log.md
- conftest.py
- description.txt
- pytest.ini
- README.md
- requirements.txt
- .gitkeep
- agent_execution.png
- agent_loop.png
- .gitkeep
- gitignore
- sandbox_bashrc
- base_requirements.txt
- Dockerfile
- GOOGLE_APPLICATION_CREDENTIALS.json
- install_swebench_harness.sh
- .gitignore
- CITATION.cff
- LICENSE
- Makefile
- README.md
- runner.py
# Installation Guide
git clone https://github.com/MaximeRobeyns/self_improving_coding_agent
Downloads the entire project code from GitHub to your computer.
cd self_improving_coding_agent
Moves into the project folder you just downloaded.
2. Official Install Script
Easy Recommended- Python 3 Python is required to use pip.
pip install swebench
Installs the package published on PyPI directly β no need to clone the source.
Pulled directly from this repo's README.
3. Python
Easypip install -r base_agent/requirements.txt
Installs the Python libraries listed in requirements.txt (or similar).
pip install swebench
Installs the package published on PyPI directly β no need to clone the source.
python -m agent_code.agent --server true -p "<some initial request here>"
Runs the Python script (or module).
python -m base_agent.agent --help
Runs the Python script (or module).
Pulled directly from this repo's README.
4. Make
Medium- Git Needed to download the project code from GitHub.
- Make Usually pre-installed on Linux/macOS. On Windows, install separately (e.g. via MSYS2 or WSL).
make image
Compiles the code based on the generated build configuration to produce an executable.
make image-mac
Compiles the code based on the generated build configuration to produce an executable.
make int
Compiles the code based on the generated build configuration to produce an executable.
make help
Compiles the code based on the generated build configuration to produce an executable.
Pulled directly from this repo's README.
