Repository navigation
Getting Started
Isaac Thor edited this page Mar 8, 2026
·
1 revision
- Python 3.10+
- Docker (optional, for containerized runs)
- Git
git clone https://github.com/codethor0/recursive-meta-learning-workbench.git
cd recursive-meta-learning-workbench
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install .For development (tests, linting):
pip install .[dev]Baseline mode runs the Web Attack Workbench without learning:
rmlw scan --target http://localhost:8080 --mode baseline --format humanWith JSON output:
rmlw scan --target http://localhost:8080 --mode baseline --output findings.jsonLearning mode uses the MetaController for adaptive testing:
rmlw scan --target http://localhost:8080 --mode learn --iterations 10 --output findings.jsonBuild:
docker build -t rmlw:local .Run (TARGET_URL required):
docker run --rm -e TARGET_URL=http://localhost:8080 rmlw:localRun tests inside the container:
docker run --rm rmlw:local pytest -v- Uncomment the
dvwaservice indocker-compose.yml - Start DVWA:
docker compose up -d dvwa- Log in at the DVWA URL (default admin/password), set security to Low
- Run RMLW:
docker compose run --rm rmlw rmlw scan --target http://dvwa --mode baseline --format human-
Docker network issues with DVWA: Ensure RMLW and DVWA are on the same Docker network. Use
docker composeso both services share the default network, or use--network hostwhen running RMLW against a host-published DVWA. - Connection refused: Verify the target URL is correct and the target is running. For DVWA, confirm you have logged in and set security level.
- No findings: Baseline mode uses a fixed set of endpoints. If your target has different paths, you may need to configure endpoints programmatically or wait for crawler support.