aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* test(storage): cover records, vault, PINs, lockout and migrationscirakg10 days7-1/+1065
| | | | | | | | | | 34 new tests (65 per device): round trips per entry type, every tampered byte, swapped records, the pending queue, the entry cap, a full store at 67 fill levels, the lockout schedule on a fake clock including restarts, erase, PIN change and reset, auto-lock wiping, and v0 to v1 migration. Adds an in-memory store with a size limit and storage benchmarks. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* feat(storage): add the encrypted vault, PIN keys, lockout, session and ↵cirakg10 days12-0/+1605
| | | | | | | | | | | | | | migrations Records are one ByteArray each: a readable TLV header plus an encrypt-then-MAC box whose aad binds the entry ID and the header. The device key seals unlocked entries and pending items; a random master key, wrapped under a PBKDF2 PIN key, seals locked ones. Wrong PINs are counted before the check, with delays on the trusted clock and an uptime fallback. Bytes.toUtf8 validates UTF-8 first, because the native conversion crashes on invalid input. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs: record M2 results and bump version to 0.2.0v0.2.0cirakg10 days7-4/+84
| | | | | | | | | CHANGELOG 0.2.0; PERFORMANCE (simulator benchmarks and what they mean for the watch); TESTING (vector sources, M2 results, watchdog caveat); FEASIBILITY (verified: no native HMAC-SHA1, AES-256 works, Long semantics); README status and targets; developer-guide Monkey C facts. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* feat(spike): add the M1 settings spike app and make targetscirakg10 days37-5/+861
| | | | | | | | | | A separate throwaway app (own app ID, no permissions) that shows what the Garmin Connect app delivers per settings field and writes values back, to measure the undocumented settings limits on the real phone and watch. Includes length-marked test strings, `make spike` / `make spike-sim`, `make vectors`, and step-by-step instructions in docs/SETTINGS_SPIKE.md. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test: add cross-checked crypto/OTP vectors, unit tests and stricter screenshotscirakg10 days14-2/+1193
| | | | | | | | | | | | | - gen_vectors.py: RFC/FIPS/NIST vectors recomputed with Python's stdlib (openssl for AES) before use, plus seeded cases around SHA padding boundaries, 2038 and 2100; generates test/data/Vectors.mc. - 31 Monkey C tests (hashes, HMAC, PBKDF2, HKDF, AES, sealed records incl. every-byte tamper detection, Base32 errors, 100 OTP and 12 Steam vectors, clock, jobs) and a benchmark; pass on all four test devices. - Screenshot tolerance 0.01 %: one changed character no longer passes. - Golden images updated for version 0.2.0. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* feat(otp): add Base32 decoding, HOTP, TOTP and Steam Guard codescirakg10 days2-0/+198
| | | | | | | | Base32 accepts any case, spaces, hyphens and missing padding, and reports the exact position of a bad character. TOTP uses 64-bit time, so codes stay correct after 2038. Steam Guard is a separate, optional code type. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* feat(crypto): add SHA-512, HMAC, PBKDF2, HKDF, AES-256 records and self-testcirakg10 days10-0/+923
| | | | | | | | | | | | | | | | - Sha512: FIPS 180-4 in Monkey C (no native SHA-512), rotations inlined with precomputed masks; constants derived from the primes, not typed. - Hmac: RFC 2104 over SHA-1/256/512 (native HMAC only supports SHA-256). - Kdf: resumable PBKDF2-HMAC-SHA256 job with calibration bounds (2,000-200,000 iterations) and HKDF-SHA256. - Aes: CBC with PKCS#7 over the native cipher; AES-256 chosen. - SealedBox: encrypt-then-MAC records (HKDF-split keys, length-prefixed aad, constant-time tag check before decryption). - SelfTest: known-answer tests for every primitive at start-up. - Bytes (wipe, constant-time equal), Clock (2038-safe widening, trusted clock), Jobs (bounded steps against the watchdog, timer runner). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs: add developer, testing and performance guides; record M0 resultsv0.1.0cirakg10 days7-8/+238
| | | | | | | | | | | | - DEVELOPER_GUIDE: workflow, how the tooling works (verified simulator behaviour), conventions and how to add devices. - TESTING and PERFORMANCE: test layers, M0 results and measured numbers. - FEASIBILITY: what M0 verified (watchdog 120k/240k, 10 MB storage in device profiles, unreliable monkeydo exit code, icon sizes). - README status, new make targets and layout; CHANGELOG 0.1.0. - Author/Summary headers on every document, including the original brief. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* feat(app): add minimal watch app with smoke test and per-size iconscirakg10 days29-0/+381
| | | | | | | | | | The M0 placeholder app: AppBase entry point, home screen with name and version, debug-only diagnostics (start-up time, memory) and a unit test that runs in the headless simulator. Builds with strict type checking and zero warnings for all 72 products; icons.jungle gives every product a launcher icon of its exact size. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(tools): add screenshot, budget and product checks; enforce file headerscirakg10 days21-81/+844
| | | | | | | | | | | | | | | | | | - sim_screenshot.py finds the device drawing in the simulator window, crops the screen and compares it with approved golden images. - check_budgets.py reads key=value DIAG/HARNESS lines (the simulator console escapes quotes, so JSON was dropped). - ciq_test_report.py accepts SDK 9.2.0's real summary line `PASSED (passed=N, failed=N, errors=N)`; a captured log is a fixture. - gen_products.py checks all manifest products are watches with API 5.0+ and at least 512 KB for apps. - lint_project.py rule HDR001: every file carries the Author/Summary header (JSON keys, PNG text chunks; generated package-lock exempt). - make_icon.py --all draws one icon per launcher size in use. - Golden screenshots for fr255s, venusq2, fr70 and fr965. - 44 tests, warnings as errors, ruff and mypy --strict clean. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* build: complete the Docker pipeline for build, test, screenshots and packagingcirakg10 days20-5/+437
| | | | | | | | | | | | | | | - Container scripts now run from the mounted project (no image rebuild per script change): build-app.sh (strict -l 3, warnings fail), run-unit-tests.sh, screenshots.sh, app-diag.sh (perf/memory), lint.sh, package.sh and sim-vnc.sh (noVNC on 127.0.0.1:6080). - Headless simulator helpers: wait for port 1234, remove the stale single-instance lock (~/Sim-dev) that made later starts hang, show the simulator log on failure. - New targets: screenshots-approve, icons; `make lock` re-adds the requirements header; pytest treats warnings as errors. - Author/Summary headers on every build file. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(tools): add test-log judge, docs and policy checks with testscirakg10 days12-0/+890
| | | | | | | | | | | | | | | - ciq_test_report.py: decides pass/fail from Run No Evil output instead of monkeydo's unreliable exit code; writes JUnit XML. - check_docs.py: offline check of relative links and GitHub anchors. - lint_project.py: project rules (headers, no logging outside Diag, no secret-like literals, no network APIs or permissions). - check_budgets.py with budgets.json: perf and memory gates. - make_icon.py: reproducible launcher icon. - 31 pytest tests, including Hypothesis fuzzing of the log parser and integration checks against the real docs and sources; ruff and mypy (strict) clean. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* build: add Docker-only Makefile, SDK setup and product manifestcirakg10 days9-0/+558
| | | | | | | | | | | | | | | | | | - Makefile: every target runs in a container as your UID/GID, with no capabilities, a read-only root filesystem, only the project folder mounted and no network except for image builds and `make sdk`. Serial even under `make -j`. `make nuke` removes containers, images, the dedicated buildx builder and .cache/ after confirmation. - sdk-setup.sh: agreement acceptance pinned to the hash you read, password login (never stored), SDK 9.2.0 and device download, and the login token deleted afterwards. - gen-key.sh: RSA-4096 developer key into .cache/keys, never overwritten, with backup instructions. - manifest.xml: watch-app, minApiLevel 5.0.0, no permissions, and the 72 watches with API 5.0+ and at least 512 KB for apps. - .gitignore, lint configs and CHANGELOG. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* build(docker): add pinned dev and docs imagescirakg10 days7-0/+4375
| | | | | | | | | | | | | | | | | garmin-otp/dev: Ubuntu 22.04 (pinned by digest; the simulator needs WebKitGTK 4.0), apt packages from a dated snapshot.ubuntu.com mirror, OpenJDK 17, Xvfb/openbox/noVNC for the headless simulator, the connect-iq-sdk-manager CLI v0.8.4 checked against its release SHA-256, and Python tools installed from a hash-locked requirements file. garmin-otp/docs: Node 22 (pinned by digest), Debian snapshot packages, markdownlint-cli2 and mermaid-cli from package-lock.json, and the headless Chrome build pinned by Puppeteer. Locks only use releases published before 2026-10-01. The Garmin SDK, device files and the developer key are never baked into an image. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs: record final decisions, test watch and SDK login detailscirakg10 days2-47/+58
| | | | | | | | | | | - README: Forerunner 70 is the test watch, the phone is Android, Docker access is settled; how `make sdk` logs in and why the browser login can't be used; the new `make products` and `make lock` targets. - FEASIBILITY: final decisions D1-D5. - Space out table separator rows and turn bold-only lines into headings so markdownlint passes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Inital commit.cirakg10 days3-0/+1480