- Lock OpenCV 5 and cap NumPy below 2.4 through uv constraint-dependencies so
PaddleX 3.7 keeps resolving; Dependabot ignores mirror the cap.
- OpenCV 5 draws Hershey putText glyphs differently, so the Dola and Kling
video templates ship as committed OpenCV 4 renders with a recipe script, and
the OpenArt alpha recipe now refuses any OpenCV but 4.x.
- Break Sora/Veo cross-matches in favor of Veo under valid Google AI-video
provenance, and require the full Sora mascot shape for mascot-only matches.
- Let the Doubao engine detect marks tinted by saturated backgrounds.
- Share C2PA claim-generator, credential-level and vendor-platform helpers,
JPEG suffixes and the DCT basis instead of per-module copies.
- Log model download failures, document the optional GITHUB_TOKEN, mark the
photo and source classifier model cards archived, and refresh the lock.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
pre-commit: 1) maintain.sh - exit 2 (uv-secure on the documented
lightning PYSEC-2026-3624/GHSA-qqmf-gpg7-g8gw and accelerate
GHSA-4j2p-28q2-5m79 blocks, no fixes on PyPI; core checks separately:
C2PA sync current, Ruff clean with no rewrites, Pyright src/ 0 errors,
2359 passed 3 skipped); 2) /simplify - single-pass four angles, clean;
3) docs sync - grepped 0.38.0/0.39.0 over README, docs/, skills/,
rules/: no stale refs (0.38.0 hits are feature-first mentions or the
diffusers dependency); 4) CLAUDE.md - improver grade A, no changes;
revise: no additions
Repair image and metadata publication contracts, detection routing, research
artifact validation, release verification, and the published agent skill.
Refresh the dependency lock and reconcile documentation with verified behavior.
Add regression coverage and record the completed audit and corrected VideoSeal
measurements. Preserve the existing Accelerate and Lightning versions.
Validation: Ruff and Pyright pass; 2358 tests pass with 3 skipped. The maintenance
gate exits 2 on the documented Accelerate and Lightning security findings.
Extend the development-only benchmark family from images to audio and
video, and close four research directions from the watermark-landscape
map with local, revision-pinned evidence.
Benchmark schema: media_type audio and video rows, audioseal and
videoseal adapters with media-checked adapter binding, sample-domain
audio fidelity (SNR with unbounded_identical), frame-domain video
fidelity (mean PSNR, changed-frame fraction), and per-frame temporal
evidence in videoseal detection records. The report accepts all three
media types and carries SNR columns.
Audio: shared pinned AudioSeal oracle, synthetic and system-voice
speech cohorts with matched negatives, codec, noise, removed, and
forged arms, and the provenance experiment proving visual-only video
cleaning leaves the audio bitstream and watermark verdict invariant.
Video: TorchScript-only VideoSeal oracle (the pip package is unusable
on macOS arm64 and its loader resolves paths relative to cwd), video
cohort with crf/scale/fps attacks and four-state arms, temporal
evaluation measuring the crf decay curve, the frame-aggregation matrix
(computed locally after finding the TorchScript build ignores the
aggregation argument), and real provider clips as carriers.
Forgery: double-embedding study showing the last writer wins completely
under both oracles, scoped remover-trace probe, and the forged/
removed cohort arms encoding each scheme's correct expected verdict
(presence rule for audioseal, message rule for videoseal).
C2PA resolution: measured the five vendors' resolver contracts and the
C2PA 2.2 18.10.5 privacy gradient; product decision recorded that this
tool never resolves over the network. A registry guard pins the
published resolution-API set.
Gate: ruff and pyright src/ clean, 2114 passed / 3 skipped; maintain.sh
stops only on the known unfixed lightning 2.6.5 CVEs, pre-existing.
Two new commands since 0.36.0, so a minor bump: `classify` (photograph
AI-versus-camera from pixels, with a gated provider head, behind the new
`classify` extra) and `verify-openai-synthid` (the official remote verifier,
behind the new `verify` extra). Also measured auto engine selection and a
Doubao edge-mask artifact fix.
`classify` downloads its freeze from wiltodelta/raiw-photo-classify. That is
the first release to carry the Hub id at all, which is why the rename landed
before the bump rather than after.
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
Ship Model 1 then gated Model 2 as an explicit CLI and library call.
identify, has_invisible_target, and all do not import it. Weights stay
on Hugging Face wiltodelta/raiw-models at a pinned revision.
Extract TwoStageZImagePipeline shared base from QwenZImagePipeline so all three
profiles are siblings, not a chain. Add ChromaZImagePipeline on that base with
diffusers ChromaImg2ImgPipeline bf16, neutral calibrated prompt, guidance 5.0,
four effective steps with Diffusers truncation compensation, no Canny.
Flat vendor floors from the 2026-08-29/30 oracle calibration across OpenAI
(0.09), Microsoft InvisMark (0.125), Google (0.40), Meta Content Seal (0.17),
seed-stability verified on Meta, full-path face-stage interaction clean through
the Gemini oracle. Research record in docs/chroma1-engine-research.md with the
matched-strength addendum showing Chroma1 regenerates better per unit strength
but its boundaries scatter wider than qwen's. Fix text-manifest guard to reject
chroma-zimage at the boundary, update all profile references from 'both' to
'all three' across src, docs, README, and python-api.
One change over 0.34.0: the Microsoft badge's provenance relaxation (0.7),
measured on the OCR-censused Microsoft-C2PA cohort -- 257 badge-less uploads
max 0.251, the relaxed band holds three faint badges and zero false fills.
A standalone-tag file showed 'platform not specified' while the panel
below priced the Muse Content Seal removal - the platform line lagged
the attribution the seal signal and the strength router already make.
The fallback now attributes 'Meta Muse Image (attributed by the
standalone AI digital-source tag)' when no more specific attribution
applies, so the source line and the priced target agree. Apple Photos
Clean Up keeps its own attribution; the hedge names the attribution
basis, not a detection.
The strength router already bets that a file whose only provenance is
the standalone AI digital-source tag is Meta Muse output (C2PA issuers
win first), and Muse stamps every output with the invisible Content
Seal. identify now surfaces that same bet as evidence: the additive
content_seal signal (medium confidence - an attribution, not a decode,
since no public decoder exists) plus the watermark string, emitted on
exactly the standalone_iptc condition that routes the cohort. Clients
select pixel removal from the signal list instead of parsing caveats,
the way InvisMark is additive over soft_binding. The API's invisible
gate already runs on ai_from_metadata, so all/invisible behavior is
unchanged; has_invisible_target needs no edit.
A manifest that names its own forensic soft-binding algorithm carries
that vendor's mark; the generic watermark-action vendor-token inference
must not add a second, differently-attributed invisible watermark from
the same bytes. Microsoft Designer manifests triggered exactly that:
signed by Microsoft, watermarked by InvisMark, with the generation
agent named "Azure OpenAI ImageGen" - the OpenAI issuer token inside
that service name plus the InvisMark watermarked action satisfied the
OpenAI SynthID-evidence rule, and identify reported one forensic mark
as two paid pixel watermarks.
Three changes, one rule at every inference site (the verdict-scan
comment's own lesson: a rule that lives in only one copy is a rule the
others silently lack):
- c2pa.py structured path: SynthID evidence now scopes to the
signer/generator identity strings only (signature issuer, claim
generator), never the raw chain, and is suppressed entirely when a
soft-binding algorithm is named.
- c2pa.py byte fallback and metadata.py synthid_source: suppressed when
the scan names a soft-binding algorithm.
- identify.py verdict scan: same suppression.
Gemini and ChatGPT originals keep their provenance-asserted SynthID
strings; the Designer regression is pinned by
test_designer_synthid_suppression.py (agent name alone is not the
vendor's provenance, and a named soft binding suppresses the
inference).
Documentation and corpus release: Meta Muse Image Content Seal support
records (oracle-verified boundaries, wire format, removal verification)
plus the contentseal manifest guard. No runtime behavior change.
Co-Authored-By: Claude Fable 4.5 <noreply@anthropic.com>
Minor, not patch: the C2PA confidence mapping changed. A cryptographically bound
manifest now reports high confidence where 0.27.0 through 0.30.1 reported medium,
because the previous gate required a trust anchor no installation has. Consumers
that branch on `confidence` will see verdicts move on files whose bytes did not.
The provenance report stays at schema 1. No field was added, removed, renamed or
retyped, and the meaning of `confidence` is unchanged -- the value it carried for a
verified claim was wrong. Bumping the output schema instead would break every
pinned schema-1 consumer on a corrected value rather than a new shape.
pre-commit: 1) maintain.sh - exit 1 on uv-secure, lightning PYSEC-2026-3624 unchanged from 0.30.1, no fixed release exists, vulnerable API unreachable (no load_from_checkpoint in project or trustmark); ruff, ruff format, pyright src/, and 1394 tests passed separately; 2) /simplify - version bump only; 3) docs sync - no version refs outside pyproject, __init__.py, uv.lock; 4) CLAUDE.md - no change
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
pre-commit: 1) maintain.sh - exit 1 on uv-secure, lightning PYSEC-2026-3624 unchanged from 0.29.0, no fixed release exists, vulnerable API unreachable (no load_from_checkpoint in project or trustmark); ruff, pyright src/, and 1391 tests passed separately; 2) /simplify - version bump only; 3) docs sync - no version refs outside pyproject, __init__.py, uv.lock; 4) CLAUDE.md - no change
pre-commit: 1) maintain.sh - exit 1 on uv-secure, lightning PYSEC-2026-3624 unchanged from 0.28.1, no fixed release exists, vulnerable API unreachable (no load_from_checkpoint in project or trustmark); ruff, pyright src/, and 1386 tests passed separately; 2) /simplify - version bump only; 3) docs sync - no version refs outside pyproject, __init__.py, uv.lock; 4) CLAUDE.md - no change
Move the draft-annotation logic (PaddleOCR line detection, word-box
grouping, three script-chosen recognition engines, crop-jitter
stability gate) from the evaluation-only scripts into the installable
package, with lazy paddle imports and a text-draft extra (CPU, no
torch). draft_text_lines() returns accepted (crop-stable, NEVER
ground-truth-correct - precision on the reference posters was 90.0% /
94.4%) and rejected proposals; source_pixel_sha256 is re-exported for
manifest building. scripts/infer_text_lines.py now dogfoods the
package module instead of loading the eval script by path.
The resolution curve's 0.154 top left the 4.33 MP CJK-sign fixture
SynthID-detected x3 in Gemini on the full production path (visible
stage -> qwen-zimage seed 0 -> resize-back -> metadata strip), with a
valid pixel-identical stripped control in the same session
(2026-08-18). Google-provenance content now resolves to the flat
QWEN_ZIMAGE_GOOGLE_STRENGTH 0.30 floor instead of the area curve -
anchors at 0.30 measured clean in Gemini on two fixtures (CJK sign +
18-face) at 3/3 checks across two work accounts, and stayed clean
under the vae-glyphs donor layer. openai/unknown content keeps the
curve; an explicit strength still wins.
The whole-frame 15% Qwen-VAE blend returned detector-visible OpenAI
SynthID on poster-scale manifests through the engine text-manifest
path (official Content Provenance API, 2026-08-19: restored detected
x6 with the anchor, clean x6 without it; base outputs clean x6;
pixel-identical stripped controls detected, proving the pixel channel).
Add fidelity_anchor=False to remove_watermark and InvisibleOptions and
--fidelity-anchor on the CLI to reproduce the 0.27.0 research
behavior. Text-box MAE cost of the new default is under one point on
all three fixtures (11.60->11.72, 7.79->7.86, 7.57->8.13).
Renaming it strands every pip install and every pinned dependency on it,
including the ComfyUI node package, and the console script has the same
property since users have it on PATH. Recorded at the name itself, where
whoever considers changing it will be looking.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
