Boundary Engine enforces claim discipline by checking public repo language against recorded proof artifacts.
U27-S07
BOUNDARY ENGINE
CLASS: SYSTEM
OPERATING_POSITION: 07/08
FUNCTION: Claim Boundary + Proof Alignment
REF_ID: U27-S07-BOUNDARY-ENGINE
SOURCE_STATUS: PUBLIC_PACKAGE
ACCESS_STATUS: CLEARED_FOR_EXTERNAL_USE
This repository is a released Unit27 field kit: visible, inspectable, and intended for orientation, testing, and practical use. Controlled protocol materials remain outside this source package.
It answers one narrow question:
Do this repo's public claims stay inside its recorded proof?
Use Boundary Engine before publishing a repo, demo, or research artifact to catch public claims that outrun recorded evidence.
It is useful when a README says things like production-ready, tested, validated, or ready for review, but the repo needs an explicit proof trail for those claims.
Example:
Claim: This repo is production-ready.
Result: unsupported unless a matching passing proof case exists.
The current public release is GitHub-first. Run it from a local checkout:
git clone https://github.com/unit27research/unit27-boundary-engine
cd unit27-boundary-engine
pip install -e .
boundary-engine demo
cat boundary-engine-demo/u27/BOUNDARY_REGISTER.mdOn your own repo:
boundary-engine scan README.md --proof u27/proof_ledger.jsonBoundary Engine scans markdown claims and writes:
u27/BOUNDARY_REGISTER.mdu27/boundary_report.json
It flags unsupported public language such as production-ready, broad proof claims, guarantee language, and review-readiness language when no matching passing Proof Ledger claim exists.
It is designed to feel like a deterministic claim-boundary check, not a legal review or copywriting assistant.
For local development:
pip install -e .Install directly from GitHub:
pipx install git+https://github.com/unit27research/unit27-boundary-engineboundary-engine scan README.md --proof u27/proof_ledger.json
boundary-engine demoExit codes:
0 = no unsupported claims
1 = unsupported claims found
2 = scan error
README -> extract claim risks -> compare proof ledger -> write boundary register
Boundary Engine does not decide whether a project is good. It checks whether public-facing language has outrun recorded evidence.
Boundary Engine is released as part of the Unit27 public tooling channel. CI verifies the test suite, wheel build, and wheel contents before changes are considered ready.
Boundary Engine does not:
- Certify truth
- Replace legal or compliance review
- Rewrite full documents
- Judge product quality
- Infer evidence that is not in the proof ledger
PYTHONPATH=src python3 -m unittest discover -s tests
boundary-engine demo
boundary-engine scan README.md --proof u27/proof_ledger.jsonPYTHONPATH=src python3 -m unittest discover -s tests
PYTHONPATH=src python3 -m boundary_engine.cli demo --root examples/sample-project
PYTHONPATH=src python3 -m boundary_engine.cli scan README.md --proof u27/proof_ledger.json --root examples/sample-project || test $? -eq 1
python3 -m pip wheel . --no-deps --no-build-isolation -w /tmp/boundary-engine-wheel
python3 scripts/verify_wheel.py /tmp/boundary-engine-wheel/unit27_boundary_engine-0.1.0-py3-none-any.whlMIT