LR-PayloadLab accepts improvements that make bounded experiments more observable, reproducible, or easier to verify.
A new action must:
- be clearly benign;
- be statically inspectable;
- have an explicit resource/path cap;
- produce an auditable receipt;
- have rollback or cleanup semantics when it creates artifacts;
- remain local-only;
- include tests for boundary enforcement.
Run:
python -m pip install -e .
python -m unittest discover -s tests
payload-lab inspect examples/telemetry-demo.jsonDo not submit features for stealth, evasion, persistence, credential access, exploitation, or remote control.