Add perceptual-hash image dedupe

[JE-Chen]

Dependency-light batch (Pillow is already core). Full layers + tests + EN/Zh v42 docs + README.

Feature (utils/image_dedup, Pillow only)

• average_hash / dhash (perceptual fingerprints via Pillow), hamming_distance / images_similar (bit distance + tolerance), dedupe_images (collapse near-duplicate frames, first wins). Maps visually similar screenshots to close fingerprints so a recording/step report's redundant frames cluster and collapse.
• No new dependency — Pillow (pillow==12.2.0) is already a core dep; perceptual hashing is implemented directly (no imagehash). The dedupe/compare logic is pure Python with an injectable hasher, so clustering is unit-tested without any image; the real Pillow hashing path runs under importorskip.
• Executor AC_image_hash / AC_dedupe_images (paths as list or builder JSON string); MCP ac_*; Builder under Image.

Verification

• 8 tests pass (hamming/similar, dedupe via injected hasher + strict threshold + empty, real Pillow hashing under importorskip, executor round-trip with generated PNGs, wiring, facade); ruff clean; radon no CC≥C; bandit clean; PySide6-free.
• Stacked on S3 (Add S3-compatible artifact store #249); rebased cleanly onto dev.

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

🟢 Metrics 33 complexity · 0 duplication

Metric	Results
Complexity	33
Duplication	0

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud