The packaging tests a catalog runs say nothing about content, so the skill's
prose drifted from the CLI exactly the way an unguarded seam does. These tests
read the live Click tree: mark keys, flag choices, batch modes, pipeline
profiles, vendor cohorts, exit codes and extras must match what the skill says.
Guard the SET, not the instances. Per-flag comparisons only check what the skill
already documents, so `--vendor`, `--pipeline auto`, and then the whole
`classify` and `verify-openai-synthid` commands each shipped with a green suite.
`test_every_cli_command_is_named_somewhere_in_the_skill` closes that.
None of it proves the only thing a skill is for: that a model reaches for it and
then obeys it. `scripts/skill_eval.py` builds a throwaway project with the skill
installed and a recording stand-in for the CLI on PATH, runs a headless agent on
one user prompt, and grades the recorded command trace. Grading is mechanical --
"did it run `all` where the probe said nothing is written" is a fact in the
trace, and "did it name an output file that is not on disk" catches a fabricated
result without asking a model to judge itself. Agents are nondeterministic, so
the report is a pass rate per check over repeats, never one verdict.
The harness is tested too: the stand-in's exit codes are compared against the
real CLI where the state is reproducible locally, the grader runs over hand-built
traces, and every case is checked for internal consistency, because a scenario
whose `identify` names a mark its `visible` cannot find measures the case rather
than the skill.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* Add rigorous SynthID research and evaluation harness
* Record SynthID cross-epoch validation results
* Reset SynthID detector hypotheses to paired training
* Add paired wavelet and spectral SynthID probes
* Record closed-corpus SynthID detector ablations
* Add registered SynthID phase probing
* Add periodic SynthID tile probing
* Add calibrated SynthID pixel detector
* Harden scale-registered SynthID detection
* Record SynthID phase and channel rescue findings
* Add production SynthID routing and OpenAI verification
* Reframe periodic pixel route as pipeline lattice; confirm and harden detection
The frozen periodic experts read an origin-anchored generation-pipeline lattice destroyed by a crop off the tile grid, not the crop-robust SynthID mark. Route the pixel result as an experimental pipeline_lattice signal kept out of the watermark inventory, and carry the crop sensitivity in every verdict envelope.
Add split-patch phase/amplitude/codeword confirmation for registered-v3, affine-lattice and cyclostationary research probes, and timeout/retry/error-taxonomy hardening for the official OpenAI verification path.
* Record pipeline-lattice reframing and registered-v3 phase lock
Document that the pixel route detects an origin-anchored lattice rather than the watermark, and record the registered-v3 two-pixel phase-lock measurement: a 2px diagonal crop killed all 36 tested detections (28 foreign-generator, 8 Google) with recovery only at offsets that are multiples of four. Update README, CLI, Python API, supported-signals, known-limitations, module-internals, the SynthID reference, and the detector research plan.
* Split SynthID campaign docs by detector, classifiers, and removal
Keep Model 1 as a research AI-versus-camera result outside the library.
Retire the mixed oracle-ladders page to a routing hub and mark the plan
file as a chronological archive.
* Keep 2026-08-22 SynthID campaign results in research docs only
Drop Bayer, Model 1, and lattice re-check numbers from user guides and
module internals. The library surface is unchanged.
* Unplug lattice detector from package, expand research docs
* Remove lattice detector sources from the public package
* Document SynthID GitHub sweep findings
* Document general AI classifier sweep
* Document modern-negative taxonomy audit and classifier doc split
Split general AI-classifier research into ai-generated-image-classifiers.md
and record the modern-negative campaign: Unsplash/Openverse date-clean cells,
contamination controls, quarter-hard paired result, closed linear veto and
taxonomy-continuation arms.
pre-commit: 1) maintain.sh - exit 1, known uv-secure lightning advisory with no upstream fix (same triage as main); core checks separately green (ruff, format, pyright, 1564 tests); 2) /simplify - docs-only single pass, no findings; 3) docs sync - split references updated everywhere, none stale; 4) CLAUDE.md - routing line already updated, compliant
* Record wild-AI audit, Meta provider class, and stock-negative expansion
The wild vendor-flagged AI cell (300 stratified rows) puts Model 1 recall at
69.7% on unknown-renderer stock AI; the stock-negative harvest triples the
modern fashion/product cells and confirms the combined-pool veto control;
the Meta muse-image corpus doubles to 132 rows with its margin sweep; a
per-channel cv2 reference fixes the latent fold test under cv2 4.10.0.
pre-commit: 1) maintain.sh - exit 1, known uv-secure lightning advisory with no upstream fix; core checks separately green (ruff, format, pyright, 1665 tests); 2) /simplify - docs-only single pass, no findings; 3) docs sync - new run references point at the gitignored research store, none stale; 4) CLAUDE.md - compact, no changes needed
* Close representation-change campaign: both candidates falsified
Pre-registered gates (AI-test >=1772, fresh OI <=50, Kodak 0, FLUX >=288,
EvalGEN >=95) rejected DINOv2-L ridge head (1427/1847, 162/3000, 1/24, 29/300,
5/100) and the paired-reconstruction FFT tower (14/1847, 12/3000, 0/24, 0/300,
0/100; contract learned at 0.949 pairwise separation but no transfer to
production generators). Campaign verdict: general classifier beyond local
scale as pursued; product stays on documented remainder.
pre-commit: 1) maintain.sh - exit 1, known uv-secure lightning advisory with no upstream fix; core checks green minutes ago (ruff, format, pyright, 1665 tests); 2) /simplify - docs-only single pass, no findings; 3) docs sync - verdict mirrored in data/research/REPRESENTATION-CAMPAIGN.md, no stale refs; 4) CLAUDE.md - no changes
* Record full failure map, H2 router falsification, and H3 process detector closure
20 measured negative domains with batch-2 (memes, whiteboards, tickets, book
pages). H2 domain router falsified: 97.9% of AI-test looks structured, every
family 94-100%. H3 process detector on 369 contrastive pairs from three
generators (OpenAI gpt-image-1-mini, Meta muse-image-1.0, Gemini flash-image)
achieves 0.909 pairwise AUC on training pairs but transfers at AUC 0.55-0.60
to production AI; human receipts score higher than any AI family. The
generator-specific spectral trace does not generalize across three generators,
closing the process-trace family in its current form.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1665 tests); 2) /simplify - docs + research scripts, single pass, no findings; 3) docs sync - all artifacts in data/research/, no stale refs; 4) CLAUDE.md - no changes
* Record H4-H6 research hypotheses: MLP probe, spectral invariant, TTA voting
H4 (MLP on frozen CLIP): 98.6% AI-test recall but 15.8% photo FPR — same
trade-off curve as quarter-hard. H5 (spectral invariant from CVPR 2025):
2.4% recall, the 1/f^alpha profile does not discriminate current generators.
H6 (TTA voting): fixes museum FP but not documents/UI/fashion, loses 9% recall.
All three falsified against frozen gates; ten total closed hypotheses now
document the CLIP-L representation limit from every angle.
Also records the provider cascade across all 20 domains (69.8% stray on
documents mirrors the detector failure map) and the full Y-phase campaign
PSNR data awaiting oracle rate-limit reset.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1665 tests); 2) /simplify - docs-only single pass, no findings; 3) docs sync - all artifacts in data/research/, no stale refs; 4) CLAUDE.md - no changes
* Fix Y-phase frequency convention, record H4-H6 and campaign results
The initial Y-phase campaign targeted FFT bins 16-32 instead of spatial
periods 16-32 px (1/32 to 1/16 cycles/px). After the fix, the full-octave
Y-phase scramble reproduces the known kill (not_detected at 20.5 dB on s1
at native 1122x1402). Partial sub-bands are visually invisible (42.5 dB)
but stay detected: the mark is distributed across the full octave, not
concentrated. Point intervention is not viable within this frequency band.
H4 MLP probe: 98.6% recall, 15.8% FPR (same trade-off wall).
H5 spectral invariant: 2.4% recall (1/f^alpha does not discriminate).
H6 TTA voting: fixes museum FP but not documents/UI/fashion.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (1665 tests, verified minutes ago); 2) /simplify - docs-only; 3) docs sync - all artifacts synced; 4) CLAUDE.md - no changes
* Add TC260 fifth class and per-class margin calibration to provider cascade
Five-head cascade (openai/google/tc260/meta/no_ai): openai 68.8%, google
76.6%, tc260 63.8%, meta 71.8% mean recall at uniform margin 0.50.
Per-class margins calibrated against 1.67% photo budget: openai 0.31,
google 0.29, tc260 0.39, meta 0.55. TC260 test recall 71.4% at calibrated
margins on 196 held-out images. Chinese-AIGC is a viable pixel class.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (verified within the hour); 2) /simplify - research script, single pass; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes
* Record universal detector v3-v8: diverse training shifts the curve
Fresh MLP on frozen CLIP embeddings with 49 negative domains (12,258 images)
and 5,043 AI positives achieves 93.0% AI-test recall with 39/49 domains at
zero false positives. Digital art FP dropped from 99.3% (Model 1) to 24.8%,
fashion from 62-71% to 12%, UI from 99.3% to 12%. Photo FPR 0.7%. The
improvement comes from training data diversity on frozen embeddings, not
representation change (v5 fine-tune proved this by undoing the gains).
Also adds .local-eval to ruff exclude so the gate covers only tracked files.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - docs + research; 3) docs sync - all artifacts in data/research/; 4) CLAUDE.md - no changes
* Add three-tier ensemble detector: DEFINITELY/POSSIBLY/LIKELY_HUMAN
Combines Model 1 and v8 into calibrated confidence tiers. DEFINITELY_AI
(both models agree) achieves 79.6% recall at 0.50% photo FPR with 46/49
domains at zero FP. The union reaches 93.3% recall at 1.80% photo FPR
with Kodak 0/24. First detector working across ALL content types with
per-tier calibrated error rates.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - docs + ensemble; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes
* Add provider v2: diversity training eliminates document leakage
Same diversity-training insight applied to the provider cascade. Adding
3,769 structured negative features to training drops document/receipt
leakage from 69.8% to 5.1%, UI from 10% to 4.3%, fashion from 5-15% to
0-3%. AI recall improves: OpenAI 84% (was 75%), Google 88.8% (was 79%).
Photo leakage 1.4%. Worst regression: memes 47.9% (was 27%).
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - provider v2; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes
* Add universal v11: full taxonomy, 136 domains, 124 at zero FP
Master data plan mapped 148 visual domains. Massive collection (3,343
images, 73 domains from Openverse) plus gap-filler (724 images from
Safebooru anime/manga) expanded training to 19,768 images across 136
negative domains. Universal v11 (MLP-512 + Model 1 ensemble):
DEFINITELY 90.3% recall at 0.50% photo FPR, ANY 95.6% recall,
124/134 domains at exactly 0% FP, worst FP 13.8% (digital art).
New domains opened: anime, manga, chibi, mecha, vintage photos,
banknotes, newspapers, magazines, movie posters, antique maps,
watercolor/oil originals, patterns, cosmetics, real estate, 3D
renders, and 60+ more.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - docs + research; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes
* Confirm representation limit: v12 fine-tune destroys structured FP despite 6x data
Fine-tuning CLIP tower on 19,518 images achieves 95.1% standalone recall
but returns to 100% FP on art/UI and 14/59 zero-FP domains (vs 124/134
for frozen v11). This is a representation property, not a data issue.
Frozen CLIP + MLP (v11) is the only working architecture. Also records
provider v3 trained on all negative domains.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (ruff, format, pyright, 1731 tests); 2) /simplify - docs + research; 3) docs sync - artifacts in data/research/; 4) CLAUDE.md - no changes
* Document v13 plateau: 128/136 zero FP, frozen-CLIP ceiling confirmed
v13 with top-plan expansion (1,169 new images) improves domain coverage
to 128/136 at zero FP and reduces fashion FP to 5.3%, but DEFINITELY
recall plateaus at 89.6% and art FP at 15.9%. The frozen CLIP + MLP
architecture has reached its ceiling.
pre-commit: 1) maintain.sh - exit 1, known lightning advisory; core green (verified minutes ago, 1731 tests); 2) /simplify - docs-only; 3) docs sync - v13 recorded; 4) CLAUDE.md - no changes
* Document the 2026-08-31 photo-classifier freeze.
Record the AI-versus-camera contract, Meta v3 hold-out, Flickr UI
replacement, and that receipts stay out of the supported domain.
* Scrub Spaces catalog details from classifier research docs
* Record receipt layout-gate probe for the universal detector
A throwaway test CA with leaf subject CN=Ideogram, Inc signs a flat PNG
carrying a trainedAlgorithmicMedia claim. The fixture exercises the
issuer-to-platform attribution end to end through the real reader and pins
the second dimension the project separates: attribution names Ideogram
while signer trust stays untrusted. Recipe documented in the fixtures
README; no keys, vendor bytes, or user uploads are committed.
detect_doubao_frame reuses the image engine's synthetic doubao alpha as its
video template, so the glyph cannot drift between registries. The policy is
seedance-style single-floor, calibrated by sequential decode over 39
TC260-confirmed Doubao corpus videos and 25 negatives: a 12-frame stable run
at 0.35 accepts 33/39 positives and 0/25 negatives (0.30 already accepted a
negative), and through the shipped arbiter identify_video selects 32/39 with
no false selection. Also documents the measured impossibility of a bare-pill
removal arm with the current detector (68 OCR-confirmed pills max 0.317 vs
clean negatives max 0.353 -- no separating threshold exists).
One committed example per registered mark: 12 PNG (image registry) and 6 MP4
clips (video registry), generated by scripts/render_visible_examples.py from
the committed silhouettes and detector templates -- never from user uploads.
The generator self-verifies (exit 1 when a mark misses its own example) and
tests/test_visible_examples.py holds both sides to it: registry completeness
both ways, per-engine detection on the canonical example, and the shipped
temporal selection accepting each clip.
Second tranche of measured-but-unregistered candidates parked under
scripts/assets/visible-mark-candidates/ with a README recording why none
ships yet (positives do not separate from clean negatives): samsung_en,
gemini_text, notebooklm, dola, mindvideo, higgsfield, jianying, capcut, zsky,
chromastudio, digenai, gendo.
Tiled diffusion was never provider-oracle calibrated with verified text
restoration: the tiled VAE donor path ran anyway and produced results no
oracle had certified. The combination is now rejected at both the
pipeline and the engine seam (ValueError with the reason), and the CLI
help no longer implies support. The invisible help is generalized and
the metadata container list corrected (MKA/OGA/Opus/AAC).
scripts/contentseal_transforms.py reproduces the deterministic crop,
resize, and JPEG variants of the Content Seal corpus from manifest.csv,
hash-verifying every output; its README gains scripts/README.md context
and new data tests. The corpus README is honest about the one crop the
daily oracle limit left unchecked, and the eval CSVs carry the updated
verdicts. The byte-scan SynthID suppression hoists its soft-binding
lookup so the guard is computed once.
Staged on top of 0.33.1; no version bump in this commit.
Full Meta Muse Image support in the invisible-removal path:
- QWEN_ZIMAGE_META_STRENGTH = 0.1: derived by the standard
worst-boundary-plus-cross-source-spread method over five oracle-bracketed
generations (data/contentseal/manifest.csv)
- Auto mode: vendor_for_strength routes a file whose only provenance is the
standalone AI IPTC trainedAlgorithmicMedia tag onto the meta cohort; C2PA
issuers win first, so Google/OpenAI/Microsoft routing is unchanged. Muse
WebP outputs place the XMP in a tail chunk, so the scan uses the shared
chunk-aware metadata.scan_head rather than a plain head read
- Explicit override: --vendor on invisible/all/batch and
InvisibleOptions.vendor name the cohort on stripped files; naming a cohort
asserts the watermark is present, so the no-signal gate treats it like
--force at both the CLI and API seams
- sdxl-zimage has no measured Meta rung: an explicit meta vendor falls to
the conservative unknown 0.25 rather than inventing one
- identify emits a Content Seal caveat pointing at the removal path
- The legacy visible 'Imagined with AI' mark stays unregistered: a dedicated
sample hunt (newsroom mockups, community posts, press screenshots, dead
imagine.meta.com, broken Wayback captures) found no pixel-verifiable
capture, and the registry rule forbids encoding a corner without one.
erase --region remains its removal path; outcome recorded in the landscape
Co-Authored-By: Claude Fable 4.5 <noreply@anthropic.com>
Muse Image (muse-image-1.0, Meta Superintelligence Labs, 2026-07-07) ships
every output with Content Seal, a proprietary invisible pixel watermark, and
no visible mark. Establish support documentation and a verified corpus:
- data/contentseal/: five own generations via the Meta Model API, every
oracle verdict recorded in manifest.csv (44 rows, settled-text protocol,
fresh-navigation variant for calibration rows)
- Oracle: meta.ai/identification web tool only; no API endpoint exists in
the Meta Model API (verified against dev.meta.ai/docs); internal REST
pair documented with its server-side sliding-window rate limit
- Removal: default qwen-zimage profile clears Content Seal (oracle-verified
on the worst source); strength floor derived at 0.1 by the standard
worst-boundary-plus-cross-source-spread method, recorded but not encoded
as a constant since no provenance signal routes Muse output onto a vendor
cohort
- Seal robustness measured: survives resize, JPEG q85, metadata stripping,
CDN WebP transcode; dies to center crops and diffusion regeneration
- tests/test_contentseal_corpus.py guards manifest integrity
Co-Authored-By: Claude Fable 4.5 <noreply@anthropic.com>
Everything here is evidence and state, not behavior: the shipped operating point
is untouched at 512 px / 12 fps / noise_std=0.15.
The carrier was downloaded and its sha256 matched the manifest, so its properties
are now measured rather than assumed: 1920x1080 at 24 fps with an AAC audio track.
Those three fields go into the two 2026-07-31 rows, which could not previously
tell a reader what the downscale factor even was.
The geometry prize is measured end to end for the first time. Against the
untouched source, 1024 scores +3.46 dB over the shipped 512, while the entire
noise_std axis is worth 1.92 dB. Two readings that the table alone hides are
recorded with it: the temporal residual IMPROVES with resolution, because the
shared noise field lives on the latent grid and is four times finer relative to
the frame at 1920; and the frame-rate arm cannot be judged by these metrics at
all, since they price its cost and not the smoothness it buys.
Two findings that outrank the quality question. The certified row does not
reproduce -- a rerun of the same configuration on a different device and dtype
gives a different hash and different metrics, and the manifest records neither.
And the pipeline copies audio byte for byte while stripping every metadata
marker, so it can emit a file our own detector calls clean with an untouched
Google-generated audio track inside it. The mechanism is proven on two carriers;
whether that audio carries a mark the verifier reads is not, and only the oracle
can say.
The six prepared oracle submissions were never run: file upload to the verifier
failed at the tool level. Their artifacts are gone with the scratch directory,
which is the intended lifecycle for generated media, and the document says what
rebuilding costs.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The shipped profile was certified by one oracle row, but only noise_std was
pinned: long_side and fps -- two thirds of what the verifier was actually shown
-- could move with a green suite. The test now derives the pin from
data/evaluations/video-synthid-oracle.csv, so a default without a certifying row
fails.
The certified profile is a perturbation-to-signal ratio, not a bare noise_std.
sd-vae-ft-mse publishes no scaling_factor key, so 0.18215 comes from the
AutoencoderKL class default under an upper-unbounded diffusers pin. The loader
now gates that value, carries it on VideoVaeRuntime, and passes it into encode
and decode so the validated value is the applied value. video_synthid_sweep.py
loads through the same function: the harness producing the certified rows was
the one path exempt from the gate it exists to feed.
psnr_db is measured against the already-resized frame and before the encoder, so
it cannot see the downscale, the decimation, or the codec, and no in-loop metric
can. scripts/video_fidelity_probe.py scores the delivered file end to end,
streaming the way the engine does and sharing its frame-selection rule rather
than copying it -- a frame-count check cannot catch a rule that reorders frames
without changing how many.
The manifest gains source geometry, vae, track, verbatim verdict and session
fields. The two 2026-07-31 rows keep them empty: they were never recorded and
are not recoverable. Verdicts now have four states, because the verifier's
unclear reading logged as not_detected is the silent regression the manifest
exists to prevent.
docs/video-synthid-quality-research.md records the research behind this: the
noise axis is worth about 2 dB and is nearly exhausted, resolution is the real
prize but is an uncertified destruction axis rather than a free win, and every
proposed autoencoder swap was refuted. First local measurements included.
Verified: engine output is byte-identical before and after the refactor on a
locally built clip, at noise_std 0.00 and 0.15.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
data/qwen_in/ground_truth.json is transcribed by vision (PaddleOCR mangled the
stylized Cyrillic), so the text metric scores variants against an accurate
reference instead of noisy OCR-vs-OCR. Re-measured text CER (controlnet vs qwen)
with this ground truth confirms qwen wins text across EN/RU/ZH: openai_1 0.385 vs
0.241, openai_2 0.341 vs 0.290, gemini_1 (ZH) 0.037 vs 0.000 (perfect Chinese even
at the higher 0.30 strength). Faces still favor controlnet. Refresh the numbers in
docs/known-limitations.md to this cleaner methodology.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- data/qwen_in/: a stable, committed set of 4 AI-generated images (OpenAI +
Google, carrying SynthID/C2PA -- same class as data/samples fixtures) used to
compare the controlnet/sdxl/qwen pipelines for fidelity. Two text-multi-script
(incl. RU/CJK), one EN poster, one face grid. README documents the set + the
ground-truth workflow. data/ is sdist-excluded so the wheel is unaffected.
- scripts/fidelity_metrics.py: switch text OCR from EasyOCR to PaddleOCR
(PP-OCRv6, higher accuracy esp. CJK, single multilingual stack); split into
`ocr` (seed a {basename: text} ground truth) and `compare` (--ground-truth for
a clean CER vs the hand-verified reference instead of noisy OCR-vs-OCR). Spatial
IoU-NMS keeps the best-scoring read per line so wrong-script models don't inject
garbage over Cyrillic/CJK.
- Oracle methodology: validate the OpenAI arm FIRST (openai.com/verify is more
accessible and the strongest Playwright/Chrome-MCP automation candidate; the
Gemini app is more manual). Recorded in CLAUDE.md + docs/synthid.md.
Ground-truth JSON (data/qwen_in/ground_truth.json) lands in a follow-up once
hand-verified.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
flux-1.png / flux-1.jpg are real Black Forest Labs FLUX.2 [pro] Playground
outputs (signed C2PA, issuer "Black Forest Labs" + trainedAlgorithmicMedia,
manifests verified to contain no personal data). flux-1.jpg is the first
committed JPEG-with-C2PA fixture, exercising the c2pa-python non-PNG reader path
end to end. Regression tests assert both attribute to "Black Forest Labs (FLUX)".
Also documents the verified finding (n=2, 2026-06-19): BFL's hosted output carries
the signed C2PA manifest but NOT the open invisible-watermark DWT-DCT (decodes to
degenerate all-ones, chance-level vs the FLUX reference) -- the open pixel mark is
dev-inference-code-optional only. So a hosted FLUX.2 image is identified by C2PA
alone, with no open-pixel fallback once C2PA is stripped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New samsung_engine.py mirrors the jimeng engine but anchors bottom-left; wired
into watermark_registry, the CLI (--mark samsung / auto), and identify
(visible_samsung, medium). visible_alpha_solve.py gains a corner=bl mode;
samsung_alpha.png solved from @f-liva's flat captures. Calibrated for the
Italian "Contenuti generati dall'AI" variant. Flat black/gray/white captures
committed, real photos gitignored. Tests + docs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Back docs/synthid.md section 2.2 with the actual test set: the per-image
oracle-verified subjects were only in a local working dir, while the doc claimed
they were recorded in data/synthid_corpus/. Ingest the key pos+cleaned pairs so
the claim holds.
- pos: openai_1/2/3 originals (gpt-image, openai-verify) + gemini_1/2/3/4
originals (Gemini app, gemini-app); all probe as C2PA-SynthID present.
- cleaned: OpenAI at strength 0.05 (openai_2 only s010 captured) + Gemini at 0.15
--max-resolution 1536; oracle: SynthID NOT detected. Metadata stripped, so no
C2PA on the cleaned rows.
- Excluded the third-party issue #14 image (pic3): oracle-verified but not
committed to the public corpus.
- docs/synthid.md 2.2: state OpenAI n=4 = 3 archived + 1 external-only.
- CLAUDE.md: drop the drift-prone "~65 MB" corpus size from the sdist note.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Visible-watermark work across all three corner-mark engines plus a committed,
reproducible alpha-build pipeline (scripts/visible_alpha_solve.py) fed by committed
solid black/gray/white captures.
- jimeng: new "即梦AI" wordmark remover (reverse-alpha + thin residual inpaint,
always NCC-aligned -- the mark re-rasterizes/jitters per image). Detect via glyph
silhouette NCC (0.45 threshold; does not cross-fire with Doubao). Registered in the
visible-mark catalog; `visible --mark jimeng` / `--mark auto`.
- doubao: fix a real production defect -- the shipped remover left a READABLE
"豆包AI生成" outline on real samples while detect() returned conf 0.0 (fooled by a
thin outline), so the test passed and the "56/56 clean" claim was detector-measured,
not visual. Root cause: under-estimated alpha + fixed-geometry-no-inpaint + tight
locate box. Rebuilt alpha (careful gray-self solve), always-align, thin inpaint,
widened locate box -> readable outline becomes faint texture-level traces.
- gemini: rebuild gemini_bg_{96,48} from our own controlled captures (validated NCC
0.9998 vs the prior third-party asset); removal re-verified clean, no behaviour change.
- tests: add textured-shift regression to both engines (guards the align-on-shift path
the Doubao defect exposed; lesson: a detector-only removal test is insufficient,
assert visual residual).
- docs: CLAUDE.md, README, capture READMEs and docstrings synced; stale
"exact/pixel-exact/56-clean" claims removed.
Also includes a SynthID label-wording clarification in identify.py/cli.py
("SynthID pixel watermark" -> "SynthID watermark, inferred from C2PA metadata").
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Apply fixes from a full-repo review (code, tests, docs).
Security / correctness:
- Clamp attacker-controlled PNG/caBX chunk lengths to the remaining file
size in metadata.py and noai/c2pa.py (a malformed length no longer drives
a multi-GB read); skipped chunks seek instead of read.
- noai/isobmff.strip_c2pa_boxes is now fail-safe on a malformed box: return
the original bytes with a warning instead of silently truncating the tail,
so metadata --remove can no longer emit a corrupt file.
- doubao_engine._fixed_alpha_map clamps the glyph box to the image (no crash
on degenerate width-vs-height).
- watermark_remover._run_region_hires gates the phaseCorrelate offset on
response and magnitude (a spurious shift no longer garbles text) and drops
the generator after a CPU fallback (no MPS/CPU device mismatch).
Robustness:
- gemini_engine, doubao_engine, region_eraser normalize grayscale and RGBA
inputs to BGR at the engine entry points.
- image_io.imwrite returns False on an unwritable path (matches cv2).
- invisible_engine guards a None imread result before use.
- trustmark_detector._decoder uses a double-checked threading lock.
- ctrlregen.tiling.tile_positions raises on overlap >= tile.
- humanizer chromatic shift no longer wraps opposite-edge pixels.
- identify OpenAI caveat keyed on the normalized vendor, not a substring.
- Remove the dead "visible --detect-threshold" CLI option.
- publish.yml verifies the release tag matches the package version.
Docs:
- README strength 0.05 to 0.10; .env.example HF_TOKEN marked optional;
doubao_capture README updated to reverse-alpha-only; CLAUDE.md synced with
the new behaviors and the batch command.
Tests: new test_security_clamp.py for the read clamp and isobmff fail-safe;
erase CLI coverage; integrity-clash rule 2 end-to-end; multi-tag EXIF
survival and cross-format strip guards; channel/size, tiling, humanizer, and
imwrite regressions. Full suite 493 passed, 2 skipped; ruff and pyright src/
clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add deterministic, CPU-only removal of the visible Doubao "豆包AI生成" mark and
a position-agnostic region eraser for any other visible watermark/logo.
- doubao_engine.py: locate (geometry, scales with width) + polarity-aware
white-top-hat glyph mask + cv2 inpaint; coverage-gated detection and a
dense-text safety guard. No GPU, ~30ms.
- region_eraser.py + `erase` command: inpaint arbitrary --region box(es).
Default cv2 backend (no deps); optional big-LaMa via onnxruntime (`lama`
extra, Carve/LaMa-ONNX, model downloaded on first use, never bundled).
- cli `visible --mark auto|gemini|doubao`: auto routes by detector confidence.
- tests for both engines; seed previously-unseeded CLI image fixtures to stop
the Doubao detector flaking on random corners.
- .gitignore: doubao_capture/{seeds,captures} scratch (alpha-map calibration).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
xAI Grok (Aurora) images carry no C2PA/SynthID/IPTC -- their only provenance
signal is an EXIF pair: ImageDescription "Signature: <base64>" + a UUID Artist.
Verified stable across 3 genuine generations (a real download previously read
as unknown / "no AI metadata").
- metadata.xai_signature(): matches the Signature blob + UUID Artist pair;
wired into has_ai_metadata, get_ai_metadata, and identify (platform
"xAI (Grok / Aurora)").
- data/samples/grok-1.jpg: real Grok fixture (neutral content; the Artist UUID
is the public image id, not PII).
- Tests: synthetic-fixture unit tests, real-sample assertion, identify
integration (322 passing).
Docs (research refresh, May 2026):
- C2PA 2.4 Durable Content Credentials (soft-binding re-discovery after the
embedded manifest is stripped).
- New AI-labeling laws, primary-source verified: EU AI Act Art 50 (2026-08-02),
South Korea AI Framework Act Art 31(3), California AB 853.
- Hedge removal claims: defeating the SynthID verifier is not forensic
invisibility (arXiv:2605.09203); cite SynthID-Image (arXiv:2510.09263).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Corpus images were gitignored (local-only). The negatives were reviewed and
cleared for publishing, so the labeled set is now committed (regular git, 65 MB
across 25 files) -- making the removal regression set reproducible and CI-able.
Corpus:
- Track data/synthid_corpus/images/ (pos 9, neg 15, cleaned 1); keep only the
synthetic refs/ calibration fills gitignored.
- Reconcile manifest.csv to the on-disk files: 117 -> 25 rows (92 dangling rows
for removed images pruned; dedup left one cleaned output, f6dd47a5).
- Rewrite the corpus README layout/policy (images committed; review every image
for private content before adding -- public repo, permanent history).
Test fixtures:
- Remove data/samples/not-ai-1/2/3 (personal iPhone photos, incl. GPS EXIF).
- Add the clean_photo conftest fixture serving a verified-negative image from
the corpus neg/ set; repoint the three "non-AI / clean photo" tests onto it
(skips if the corpus is absent).
Metadata-source coverage (close the last sub-variant gaps):
- c2pa digitalSourceType: algorithmicMedia (procedural, not flagged AI) and
compositeWithTrainedAlgorithmicMedia (AI + SynthID proxy).
- exif_generator: EXIF Artist and ImageDescription fields (Software/Make/XMP
CreatorTool were already covered).
All 8 metadata-source kinds are now tested at both the unit and identify()
level. 313 tests pass. CLAUDE.md updated (corpus tracked, clean_photo fixture).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add manifest row for the 4ef377bd -> f6dd47a5 chain: a gpt-image-2 sample
(openai.com/verify: SynthID + C2PA detected) cleaned via v0.5.3 `all` at
native 1254x1254 (prod-equivalent SDXL base, strength 0.05, 50 steps).
openai.com/verify reports SynthID NOT detected after the run, re-confirming
that the #10 native-resolution default defeats OpenAI SynthID and resolving
the #15 root cause (older SD-1.5/768px downscale default did not).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2048x2048 PNG carrying China's TC260 <TC260:AIGC> label; identify reports
it as a China AIGC-labeled generator (TC260). Reference fixture for manual
re-verification of the TC260 detection path -- the automated tests use
synthetic blobs, so nothing depends on this file being present.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Threat model: replace the unverified deployment list (Gemini 3 Pro /
Nano Banana Pro / Imagen 4 / Veo) with the source-verified scope -- SynthID
across Imagen / Veo / Lyria plus Gemini app outputs (>10B items by Dec 2025),
and attribute the 136-bit payload to the paper's SynthID-O variant.
openai-images-2 sample: note the file predates the 19 May 2026 SynthID
rollout across ChatGPT / Codex / API, and that openai.com/verify is now the
public oracle (still no local decoder).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- fal's llms.txt confirms fast-sdxl is stabilityai/stable-diffusion-xl-base-1.0,
the exact checkpoint the local CLI defaults to -> local == prod weights.
Recorded in CLAUDE.md and README.
- README How it works + sample README: replace the old downscale->upscale
description with native-resolution processing (matches the #10 fix);
document --max-resolution as an opt-in OOM cap.
- README roadmap: idna already bumped (uv-secure clean).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Collected live samples from three popular generators we lacked:
- Ideogram tags its downloads with EXIF Make="Ideogram AI" (no C2PA, no
SynthID, no imwatermark) -- the Make tag is its only signal. exif_generator
only read Software/Artist/ImageDescription, so it missed this; now reads
Make too. Real cameras put "Apple"/"Canon" in Make (no AI token), so this
stays low-false-positive. 4 originals ingested.
- Recraft (PNG export) and Krea hosting FLUX 2: downloads carry NO detectable
signal -- no C2PA/EXIF/IPTC, and notably no imwatermark despite Krea running
FLUX. identify correctly reports 'unknown'. Both ingested as neg fixtures.
Lesson recorded in CLAUDE.md: the imwatermark detector fires only on pristine
output from a pipeline that runs the encoder (diffusers default, official BFL),
not from re-hosts (Krea/Stability) or re-encoded exports (Recraft/Canva).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Collected live C2PA positives from Bing Image Creator and Stability Brand
Studio (DreamStudio successor) and learned two things our scan got wrong:
- Bing now runs Microsoft's own MAI-Image model, not DALL-E, and signs
C2PA as 'Microsoft'. The scan caught it, but the platform label claimed
'Microsoft Designer (DALL-E / OpenAI backend)'. Relabeled model-neutral:
'Microsoft (Bing Image Creator / Designer)'.
- Stability signs C2PA as 'Stability AI' (cert 'Stability AI Ltd'), which
was not in C2PA_ISSUERS, so it read as 'unknown signer'. Added the issuer
and a platform mapping. Stability uses no SynthID and (on its current
Stable Image model) no imwatermark watermark -- verified, both negative.
Both ingested as SynthID-negative corpus fixtures (they are AI but not
SynthID) for issuer-coverage. Canva skipped: its downloads are re-encoded
design exports that strip C2PA, so a Canva sample would be inconclusive.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Grow the SynthID corpus to 109 originals (91 iPhone-photo negatives,
2 positives) and document what was learned studying 8 platforms:
- README: per-platform watermark map (C2PA issuer / SynthID pixel / IPTC
/ visible sparkle per platform) and an "originals, not previews" note
(re-encoded previews strip metadata, so a clean preview is not proof).
- CLAUDE.md: surface-dependent blind spot -- the same Google model wraps
C2PA in the Gemini app but emits the SynthID pixel watermark + sparkle
with no C2PA/IPTC via the API/playground (AI Studio, Nano Banana), so
synthid_source returns None despite SynthID being present; only the
pixel oracle or the visible-sparkle detector catches those.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds content positives (OpenAI gpt-image: forest, fisherman, tokyo; Google
gemini: fisherman, mug) and SDXL/non-SynthID negatives to the local corpus
manifest. Now spans 4 resolutions across 2 vendors (was solid-black only).
README: documents driving generation via Chrome MCP -- Gemini single-click
download; ChatGPT via in-page fetch+blob (preserves original C2PA bytes,
unlike the flaky UI download / a canvas re-encode).
Images stay gitignored; only the manifest (sha256 + labels + extracted
metadata) and protocol are tracked.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Detect SynthID-bearing images via their C2PA companion: a manifest signed by a
SynthID-using vendor (Google/OpenAI) on AI-generated content implies an
invisible SynthID pixel watermark. Verified end-to-end against the vendor
oracles (openai.com/verify, Gemini "Verify with SynthID").
- metadata: synthid_source() + synthid_watermark verdict in get_ai_metadata,
surfaced as a `metadata --check` callout. Format-agnostic (PNG caBX parser +
JPEG/WebP/AVIF/HEIF/JXL binary scan).
- constants: SYNTHID_C2PA_ISSUERS {Google, OpenAI}; +opened/placed actions.
- c2pa: single CBOR-aware parser (_cbor_text_after) replaces glitchy regex
(fixes fGPT-4o claim_generator); removed duplicate _scan_png_c2pa_chunk from
metadata; shared synthid_verdict / synthid_vendors_in helpers.
- corpus: scripts/synthid_corpus.py ingest tool + data/synthid_corpus/
(manifest tracked, images gitignored) for a labeled reference set.
- tests: +38 across C2PA parser internals, extract/inject round-trip, ISOBMFF
container stripping, all IPTC AI markers, and invisible watermark strength
tiers (SynthID/StableSignature/TreeRing/StegaStamp/RingID/RivaGAN/...).
Pixel-level SynthID detection remains out of reach locally (Google's decoder is
proprietary); a from-scratch spectral pilot confirmed it does not separate real
content. See CLAUDE.md for the full evaluation.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- metadata --check now shows claim_generator, c2pa_spec, digital_source_type,
c2pa_actions, signer instead of empty table for C2PA-only files
- reuses existing extract_c2pa_chunk() from noai/c2pa.py — no more duplicate
PNG chunk parsing or full-file reads
- adds data/samples/openai-images-2/amur-leopard.png: real gpt-image-2 output
with C2PA manifest signed by OpenAI OpCo LLC / Trufo CA (spec 2.2.0)
- removes stale data/samples/nano-banana-1/2.png (no longer referenced)
- updates README: new Images 2.0 row in supported models table
- documents known text-degradation limitation in CLAUDE.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Rewrite README for SEO: Nano Banana, SynthID, Made with AI, C2PA keywords
- Add Supported Models table with 7 AI services
- Add 'Made with AI' label removal to features
- Rename sections for search discoverability
- Add samples: ChatGPT/DALL-E, Midjourney, Adobe Firefly
- Reorganize data/samples with flat structure and clear naming
- CLI with visible, invisible, all, metadata, and batch commands
- Gemini watermark removal via reverse alpha blending
- Invisible watermark removal via diffusion regeneration (SynthID, TreeRing)
- AI metadata stripping (EXIF, PNG text, C2PA)
- Face protection (YOLO/Haar) and analog humanizer
- 137 tests covering all CLI modes and core engines
- Ruff and Pyright clean
