5 Commits
Author SHA1 Message Date
james.maesandClaude Sonnet 5.5 089f57f759 Revert "ci: first run on ci-workflows v1 (native workflows)"
android / android (push) Failing after 3m29s
build / android (push) Failing after 3m29s
Commit 1d3847d deleted the entire tree (2584 files) and added nothing, so main was empty. This restores main to dc43a6a (the PR 22 merge) exactly.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-11-09 06:18:45 -06:00
James MaesandClaude Opus 5.5 1d3847d070 ci: first run on ci-workflows v1 (native workflows)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-08 08:26:19 -06:00
james.maesandCodex 625a275de3 fix(quality): apply client standards review and add Query About
Preserve ownership notices, replace wildcard imports, document source contracts, enforce source checks and strict Kotlin warnings, and apply the repository native formatter policy. Correct typed widget state and callback registration with regression coverage.

Add bundled Query artwork, public Dogfooding and Query links, version and license access under Settings About. Validation:163 JVM cases,46 TC22 sample native cases, QA compilation, APK builds and lint0errors/0warnings. Full formatter CI and actual external-provider Android acceptance remain explicit.

Co-Authored-By: Codex <noreply@openai.com>
2026-11-01 11:38:09 -06:00
james.maesandOpenAI GPT-6.1 6845abc528 test: select explicit scan-probe package targets
Map accepted control values to compile-time QA/E2E package literals
before constructing broadcasts. Unknown targets still return before any
send, and marker validation/actions/permissions/counts remain unchanged.
This makes the existing allowlist boundary explicit to static analysis
without suppressing CodeQL or forwarding the caller's target string.

Probe build/lint and the distinct-signer ordinary-UID device acceptance
pass: eight forged broadcasts create no extra process steps or records,
and the manual positive control saves. Production APKs are unchanged.
Static triage and fresh hosted analysis remain separate evidence.

Co-Authored-By: OpenAI GPT-6.1 (Codex) <noreply@openai.com>
Signed-off-by: James Maes <james@kof22.com>
2026-10-31 20:50:24 -05:00
james.maesandOpenAI GPT-6.1 d316567fdc feat: secure Zebra scanning and unify mobile Back routing
Use the optional device-provided EMDK service for normal scanner input,
with exact compile-only SDK provenance and no shipped vendor classes.
Keep synthetic scan broadcasts confined to the disposable E2E variant.
Pause input behind overlays and lost focus, and share live drawer-first
Back routing across every process/application handler and toolbar arrow.

Replace inherited Auth0/Coil prereleases with reviewed stable versions.
Invalidate standard generated notices on dependency graph changes and
retain strict runtime/license checks. Pin hosted acceptance to the
source-qualified API 1.9 core/server tuple with runtime provenance.

Validation: 127 JVM tests, 34 TC22 shared native cases, eight-overlay
native regression, and ordinary-UID/distinct-signer injection rejection
pass. Eight forged broadcasts cause no process/record mutation; a real
manual positive control saves through the API. All package/lint gates,
four-APK notice/DEX/native/manifest inspection, focused independent
source review, workflow syntax and staged secret checks pass.

Actual hardware triggers, configured provider, optional files deployment
and final signing/distribution remain separate documented release gates.
No normal operator installation or data was changed by these tests.

Co-Authored-By: OpenAI GPT-6.1 (Codex) <noreply@openai.com>
Signed-off-by: James Maes <james@kof22.com>
2026-10-31 20:18:41 -05:00