Audit / Validate
Record bounded evidence, then check artifacts against versioned contracts.
edgeloom audit ARTIFACT
↗
Active open sourcev0.2.0
Inspect edge drivers. Change them with a way back.
EdgeLoom audits, validates, patches, restores, translates, and discovers smart-home edge-driver artifacts.
- Apache-2.0
- Python 3.11+
- On PyPI
$ pip install edgeloom
checked · reversibleBuilt in public for developers, maintainers, smart-home users, manufacturers, integrators, researchers, and standards communities.
The boundary problem
A device can expose more than its platform profile admits.
Platform drivers translate protocol behavior into usable capabilities. When their artifacts are incomplete or opaque, attributes disappear at the boundary. EdgeLoom makes that layer inspectable and changeable.
- 01Device behaviorprotocol state
- 02Driver artifactprofiles + mappings
- 03Platform surfaceusable capabilities
Current scope: SmartThings Edge artifacts and a Home Assistant-to-SmartThings translation path — not firmware modification, hub installation, or a replacement for either platform.
Core capabilities
Six commands. Four focused workflow families.
EdgeLoom leaves packaging, review, and hub installation with the operator.
Patch / Restore
Expose mapped attributes with backups, rollback on failure, and a path to stock.
edgeloom restore DRIVER
↗
Translate
Generate SmartThings Edge proxy artifacts from supported Home Assistant entities.
edgeloom translate …
↗
Discover
Catalog driver fingerprints, devices, and gaps in capability mapping.
edgeloom discover …
↗
driver · entities · catalog→
patch · translate · discover→
schema / v0.1validate
Device evidence catalog · early pilot
Find a device.
Follow the evidence.
Compare Yale lock evidence across platform artifacts: YRD156 configuration paths and YRD210 Zigbee battery adaptation. Two identities; candidate records, no hardware evidence.
Project health
Active means inspectable.
Verified
Releases3 publicthrough v0.2.0
PackagePyPI 0.2.0wheel + sdist
Current mainCI passing3.11 · 3.12 · package · site
ContributionsPublic recordauthorship + review
SecurityGHSA responsefixed in v0.1.1
StewardshipOpen governanceroles + decisions
Activity is not adoption: no production scale, partnership, endorsement, or user-count claims are made.
Industry + standards potential
Concrete artifacts make better cross-platform conversations.
Build
Manufacturers, maintainers, and integrators can inspect how behavior becomes a capability and review generated changes.
Study
Researchers can investigate hidden attributes, trust boundaries, generated code, and driver supply chains.
Align
Standards communities can use schemas and capability maps to discuss namespaces, portability, and future contracts.
These are directions for collaboration, not claims of adoption or endorsement.
Getting started
Install. Check. Then decide.
EdgeLoom requires Python 3.11 or newer.
# install the latest release
$ pip install edgeloom
# check an artifact tree
$ edgeloom validate .
Pre-1.0: review diffs, keep backups, treat inputs as untrusted, and keep tokens out of logs.
Trust model →
Community
Use it, shape it, steward it.
Code, docs, device results, issue triage, design discussion, and review all count as contributions.
Acknowledgements
Credit belongs with the work.
EdgeLoom is maintained and strengthened in public by people who ship code, tests, documentation, security improvements, and project stewardship.
-
Chenglong Fu@infinitywings
Lead maintenance, releases, security stewardship, and integration.
Maintainer ↗ -
Victor@vsolano9
Validation and discovery fixes, translator tests, and patching documentation.
Contributor ↗ -
@LZhWangrestore workflow
The restore helper and its first-class CLI workflow.
Contributor ↗ -
@Georgefifthdiscovery + translator
Discovery limit fixes and safer Home Assistant token configuration.
Contributor ↗ -
@bluo-kuschema validation
Duplicate-entry rejection for profile contracts.
Contributor ↗ -
@HavocFiXercapability maps
Humidity-sensor capability-map coverage.
Contributor ↗ -
@chuxiongwuearly contributor
An early code contribution preserved in the project history.
Contributor ↗
Based on commits and merged pull requests in current main, verified .
Contributor graph
Merged pull requests
Research → infrastructure
A finding becomes useful when the infrastructure outlasts the paper.
EdgeLoom grows the hidden-attribute research into shared contracts, reproducible transforms, real-device evidence, and open stewardship.
The study found 119 hidden attributes across 31 devices from 16 manufacturers. It motivates EdgeLoom; it does not define current tool coverage.
Read the paper → Xu et al. · reproduced under CC BY 4.0ResearchMeasure driver/profile mismatches.
ImpactMake hidden behavior reviewable.
SustainabilityVersion, release, govern, repeat.
Preferred citation · ACM CCS 2025
Discovering and Exploiting IoT Device Hidden Attributes: A New Vulnerability in Smart Homes
Xuening Xu · Chenglong Fu · Xiaojiang Du · Bo Luo
Releases + news