Not in the top 1,000 by GitHub stars
TIGER-AI-Lab/RewardHarness
Self-evolving agentic reward framework for image-editing evaluation — 47.4% on EditReward-Bench from only 100 preference demos, no reward-model training. arXiv 2605.08703.
History
Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.
View weekly changes
examples/seed_library/skills/instruction-following-completeness/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
examples/seed_library/skills/realism-and-artifact-penalties/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
examples/seed_library/tools/text-and-ocr-analyzer/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v1/skills/evaluating-background-changes/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v1/skills/evaluating-multi-step-edits/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v1/skills/evaluating-object-replacements/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v1/skills/evaluating-style-transfers/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v1/skills/evaluating-text-rendering/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v1/tools/compare_images/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/artifact-and-structure-preservation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/artistic-and-cultural-references/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/embrace-intended-drastic-changes/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/enforce-logical-consistency/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/entity-definition-activation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/environmental-blending-and-lighting-adaptation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/evaluate-insertions-and-anatomical-integrity/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/evaluate-physical-plausibility/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/evaluate-text-additions/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/holistic-evaluation-vs-tunnel-vision/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/instruction-over-naturalness/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/mitigate-positional-bias/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/object-replacement-proportions-and-exaggeration/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/overcoming-default-to-a-bias/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/penalize-gibberish-text/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/penalize-unintended-style-changes/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/preserve-subject-interaction-and-pose/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/prevent-image-feature-swapping/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/reasoning-preference-alignment/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/strict-semantic-and-conceptual-alignment/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/subject-vs-setting-modifications/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/verify-extreme-claims-and-hallucinations/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/skills/verify-instruction-target-and-spatial-relations/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/tools/object-locator-and-counter/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/tools/side-by-side-image-comparator/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/tools/text-and-spelling-checker/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/tools/visual-detail-and-setting-analyzer/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v2/tools/visual-question-answering/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v3/skills/evaluating-multiple-constraints/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v3/skills/evaluating-style-transfers/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v3/skills/evaluating-text-additions/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v3/tools/image-difference-checker/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v3/tools/visual-question-answering/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/artist-and-style-knowledge/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/artistic-style-transfer-evaluation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/background-and-scene-transformation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/counting-and-spatial-reasoning/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/direct-vision-reliance/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/exclusivity-and-preservation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/multi-attribute-verification/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/multi-part-instruction-evaluation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/object-extraction-evaluation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/scene-and-atmosphere-transformation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/spatial-and-relational-edits/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v4/skills/text-rendering-evaluation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-broken-router/skills/mandatory-ocr-for-text/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-broken-router/skills/reasoning-preference-alignment/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-broken-router/skills/world-knowledge-and-style-awareness/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-broken-router/tools/text-and-ocr-analyzer/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-old/skills/object-attribute-verification/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-old/skills/spatial-and-relational-reasoning/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-old/skills/text-rendering-evaluation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-old/skills/thematic-and-style-transformation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-old/skills/vqa-query-best-practices/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5-old/tools/visual-question-answering/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5/skills/penalize-over-editing/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5/skills/structure-and-subject-preservation/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5/skills/verify-image-content-and-ocr/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5/skills/world-knowledge-and-entities/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5/tools/text-and-ocr-verifier/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
libraries/archive/gemini-v5/tools/visual-content-verifier/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
rewardharness/resources/library/skills/evaluating-thematic-and-setting-accuracy/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
rewardharness/resources/library/skills/score-calibration/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
rewardharness/resources/library/skills/verifying-multi-attribute-objects/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
rewardharness/resources/library/tools/fine-grained-object-verifier/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
rewardharness/resources/library/tools/image-quality-and-artifact-analyzer/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
rewardharness/resources/library/tools/visual-text-and-layout-analyzer/SKILL.md
Week of Nov 2, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → 0c1e7051ab4f
No earlier version was observed for comparison.
Verified commit 0c1e7051ab4fbc39614c7bae3bd928de0e695297
git clone --filter=blob:none --no-checkout https://github.com/TIGER-AI-Lab/RewardHarness.git .hotskills/tiger-ai-lab-rewardharness && git -C .hotskills/tiger-ai-lab-rewardharness checkout --detach 0c1e7051ab4fbc39614c7bae3bd928de0e695297
# Copy each selected skill directory from .hotskills/tiger-ai-lab-rewardharness into .claude/skills/<skill-name>/Heat
Insufficient historyNot enough history yet to measure heat. It appears after a few days of data.
Heat never changes the order by stars, and sponsorship never touches it. How heat works
Stars over time
30-day change not yet available: not enough history.
Individual skills in this repository
- instruction-following-completeness
examples/seed_library/skills/instruction-following-completeness - realism-and-artifact-penalties
examples/seed_library/skills/realism-and-artifact-penalties - text-and-ocr-analyzer
examples/seed_library/tools/text-and-ocr-analyzer - evaluating-background-changes
libraries/archive/gemini-v1/skills/evaluating-background-changes - evaluating-multi-step-edits
libraries/archive/gemini-v1/skills/evaluating-multi-step-edits - evaluating-object-replacements
libraries/archive/gemini-v1/skills/evaluating-object-replacements - evaluating-style-transfers
libraries/archive/gemini-v1/skills/evaluating-style-transfers - evaluating-text-rendering
libraries/archive/gemini-v1/skills/evaluating-text-rendering - compare_images
libraries/archive/gemini-v1/tools/compare_images - artifact-and-structure-preservation
libraries/archive/gemini-v2/skills/artifact-and-structure-preservation - artistic-and-cultural-references
libraries/archive/gemini-v2/skills/artistic-and-cultural-references - embrace-intended-drastic-changes
libraries/archive/gemini-v2/skills/embrace-intended-drastic-changes - enforce-logical-consistency
libraries/archive/gemini-v2/skills/enforce-logical-consistency - entity-definition-activation
libraries/archive/gemini-v2/skills/entity-definition-activation - environmental-blending-and-lighting-adaptation
libraries/archive/gemini-v2/skills/environmental-blending-and-lighting-adaptation - evaluate-insertions-and-anatomical-integrity
libraries/archive/gemini-v2/skills/evaluate-insertions-and-anatomical-integrity - evaluate-physical-plausibility
libraries/archive/gemini-v2/skills/evaluate-physical-plausibility - evaluate-text-additions
libraries/archive/gemini-v2/skills/evaluate-text-additions - holistic-evaluation-vs-tunnel-vision
libraries/archive/gemini-v2/skills/holistic-evaluation-vs-tunnel-vision - instruction-over-naturalness
libraries/archive/gemini-v2/skills/instruction-over-naturalness - mitigate-positional-bias
libraries/archive/gemini-v2/skills/mitigate-positional-bias - object-replacement-proportions-and-exaggeration
libraries/archive/gemini-v2/skills/object-replacement-proportions-and-exaggeration - overcoming-default-to-a-bias
libraries/archive/gemini-v2/skills/overcoming-default-to-a-bias - penalize-gibberish-text
libraries/archive/gemini-v2/skills/penalize-gibberish-text - penalize-unintended-style-changes
libraries/archive/gemini-v2/skills/penalize-unintended-style-changes - preserve-subject-interaction-and-pose
libraries/archive/gemini-v2/skills/preserve-subject-interaction-and-pose - prevent-image-feature-swapping
libraries/archive/gemini-v2/skills/prevent-image-feature-swapping - reasoning-preference-alignment
libraries/archive/gemini-v2/skills/reasoning-preference-alignment - strict-semantic-and-conceptual-alignment
libraries/archive/gemini-v2/skills/strict-semantic-and-conceptual-alignment - subject-vs-setting-modifications
libraries/archive/gemini-v2/skills/subject-vs-setting-modifications - verify-extreme-claims-and-hallucinations
libraries/archive/gemini-v2/skills/verify-extreme-claims-and-hallucinations - verify-instruction-target-and-spatial-relations
libraries/archive/gemini-v2/skills/verify-instruction-target-and-spatial-relations - object-locator-and-counter
libraries/archive/gemini-v2/tools/object-locator-and-counter - side-by-side-image-comparator
libraries/archive/gemini-v2/tools/side-by-side-image-comparator - text-and-spelling-checker
libraries/archive/gemini-v2/tools/text-and-spelling-checker - visual-detail-and-setting-analyzer
libraries/archive/gemini-v2/tools/visual-detail-and-setting-analyzer - visual-question-answering
libraries/archive/gemini-v2/tools/visual-question-answering - evaluating-multiple-constraints
libraries/archive/gemini-v3/skills/evaluating-multiple-constraints - evaluating-style-transfers
libraries/archive/gemini-v3/skills/evaluating-style-transfers - evaluating-text-additions
libraries/archive/gemini-v3/skills/evaluating-text-additions - image-difference-checker
libraries/archive/gemini-v3/tools/image-difference-checker - visual-question-answering
libraries/archive/gemini-v3/tools/visual-question-answering - artist-and-style-knowledge
libraries/archive/gemini-v4/skills/artist-and-style-knowledge - artistic-style-transfer-evaluation
libraries/archive/gemini-v4/skills/artistic-style-transfer-evaluation - background-and-scene-transformation
libraries/archive/gemini-v4/skills/background-and-scene-transformation - counting-and-spatial-reasoning
libraries/archive/gemini-v4/skills/counting-and-spatial-reasoning - direct-vision-reliance
libraries/archive/gemini-v4/skills/direct-vision-reliance - exclusivity-and-preservation
libraries/archive/gemini-v4/skills/exclusivity-and-preservation - multi-attribute-verification
libraries/archive/gemini-v4/skills/multi-attribute-verification - multi-part-instruction-evaluation
libraries/archive/gemini-v4/skills/multi-part-instruction-evaluation - object-extraction-evaluation
libraries/archive/gemini-v4/skills/object-extraction-evaluation - scene-and-atmosphere-transformation
libraries/archive/gemini-v4/skills/scene-and-atmosphere-transformation - spatial-and-relational-edits
libraries/archive/gemini-v4/skills/spatial-and-relational-edits - text-rendering-evaluation
libraries/archive/gemini-v4/skills/text-rendering-evaluation - mandatory-ocr-for-text
libraries/archive/gemini-v5-broken-router/skills/mandatory-ocr-for-text - reasoning-preference-alignment
libraries/archive/gemini-v5-broken-router/skills/reasoning-preference-alignment - world-knowledge-and-style-awareness
libraries/archive/gemini-v5-broken-router/skills/world-knowledge-and-style-awareness - text-and-ocr-analyzer
libraries/archive/gemini-v5-broken-router/tools/text-and-ocr-analyzer - object-attribute-verification
libraries/archive/gemini-v5-old/skills/object-attribute-verification - spatial-and-relational-reasoning
libraries/archive/gemini-v5-old/skills/spatial-and-relational-reasoning - text-rendering-evaluation
libraries/archive/gemini-v5-old/skills/text-rendering-evaluation - thematic-and-style-transformation
libraries/archive/gemini-v5-old/skills/thematic-and-style-transformation - vqa-query-best-practices
libraries/archive/gemini-v5-old/skills/vqa-query-best-practices - visual-question-answering
libraries/archive/gemini-v5-old/tools/visual-question-answering - penalize-over-editing
libraries/archive/gemini-v5/skills/penalize-over-editing - structure-and-subject-preservation
libraries/archive/gemini-v5/skills/structure-and-subject-preservation - verify-image-content-and-ocr
libraries/archive/gemini-v5/skills/verify-image-content-and-ocr - world-knowledge-and-entities
libraries/archive/gemini-v5/skills/world-knowledge-and-entities - text-and-ocr-verifier
libraries/archive/gemini-v5/tools/text-and-ocr-verifier - visual-content-verifier
libraries/archive/gemini-v5/tools/visual-content-verifier - evaluating-thematic-and-setting-accuracy
rewardharness/resources/library/skills/evaluating-thematic-and-setting-accuracy - score-calibration
rewardharness/resources/library/skills/score-calibration - verifying-multi-attribute-objects
rewardharness/resources/library/skills/verifying-multi-attribute-objects - fine-grained-object-verifier
rewardharness/resources/library/tools/fine-grained-object-verifier - image-quality-and-artifact-analyzer
rewardharness/resources/library/tools/image-quality-and-artifact-analyzer - visual-text-and-layout-analyzer
rewardharness/resources/library/tools/visual-text-and-layout-analyzer
How it works
instruction-following-completeness
When your agent uses it
Score completeness of instruction execution. Partial edits get partial credit; missed sub-steps cost points.
What it loads into context
Frontmatter
42 tokensAlways available to the agent
SKILL.md
354 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decomposition
- Scoring (per candidate, 1–4)
- When in doubt
Files
SKILL.mdcore1,549 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubrealism-and-artifact-penalties
When your agent uses it
Guidance on penalizing artifacts while allowing conceptual unrealism if requested by the prompt.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
279 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,294 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubtext-and-ocr-analyzer
When your agent uses it
Extracts and analyzes text within images to verify spelling, placement, and clarity.
What it loads into context
Frontmatter
251 tokensAlways available to the agent
SKILL.md
461 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- When to use
- Example query
- Reading the output
- Cost
Files
SKILL.mdcore1,859 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-background-changes
When your agent uses it
Guidelines for evaluating background replacements and scene changes.
What it loads into context
Frontmatter
28 tokensAlways available to the agent
SKILL.md
193 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore889 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-multi-step-edits
When your agent uses it
Guidelines for evaluating prompts with multiple distinct instructions.
What it loads into context
Frontmatter
30 tokensAlways available to the agent
SKILL.md
209 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore945 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-object-replacements
When your agent uses it
Guidelines for evaluating object replacements and modifications.
What it loads into context
Frontmatter
27 tokensAlways available to the agent
SKILL.md
196 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore869 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-style-transfers
When your agent uses it
Guidelines for evaluating artistic style transfers.
What it loads into context
Frontmatter
26 tokensAlways available to the agent
SKILL.md
208 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore960 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-text-rendering
When your agent uses it
Guidelines for evaluating text additions and modifications in images.
What it loads into context
Frontmatter
29 tokensAlways available to the agent
SKILL.md
206 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore863 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubcompare_images
When your agent uses it
Compare a source image and edited images to identify differences, check if instructions were followed, and assess visual quality.
What it loads into context
Frontmatter
113 tokensAlways available to the agent
SKILL.md
229 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore968 bytesSKILL.mdcore968 bytesSKILL.mdcore968 bytesSKILL.mdcore968 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubartifact-and-structure-preservation
When your agent uses it
Rules for prioritizing image integrity and penalizing severe artifacts or loss of original structure.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
227 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,084 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubartistic-and-cultural-references
When your agent uses it
Guidelines for evaluating prompts that reference specific artists, art styles, or cultural phenomena.
What it loads into context
Frontmatter
38 tokensAlways available to the agent
SKILL.md
230 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,055 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubembrace-intended-drastic-changes
When your agent uses it
Rules for evaluating setting and style changes, strictly forbidding penalizing an image for losing original context when a new setting is requested.
What it loads into context
Frontmatter
50 tokensAlways available to the agent
SKILL.md
240 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,105 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubenforce-logical-consistency
When your agent uses it
Ensures the reasoning chain does not contain internal contradictions between stated facts and image evaluations.
What it loads into context
Frontmatter
39 tokensAlways available to the agent
SKILL.md
240 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,120 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubentity-definition-activation
When your agent uses it
Requires explicitly defining specific artists, styles, or entities and their visual hallmarks before evaluating.
What it loads into context
Frontmatter
39 tokensAlways available to the agent
SKILL.md
206 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore919 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubenvironmental-blending-and-lighting-adaptation
When your agent uses it
Rules for evaluating background changes, rewarding necessary adaptations to the subject's lighting, shadows, and ground contact.
What it loads into context
Frontmatter
46 tokensAlways available to the agent
SKILL.md
348 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Guidelines
Files
SKILL.mdcore1,711 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluate-insertions-and-anatomical-integrity
When your agent uses it
Guidelines for evaluating inserted humans, animals, or objects, prioritizing anatomical correctness over superficial integration.
What it loads into context
Frontmatter
46 tokensAlways available to the agent
SKILL.md
251 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,168 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluate-physical-plausibility
When your agent uses it
Guidelines for evaluating inserted or replaced objects for physical plausibility, penalizing floating objects or gravity-defying placements.
What it loads into context
Frontmatter
44 tokensAlways available to the agent
SKILL.md
229 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,097 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluate-text-additions
When your agent uses it
Guidelines for evaluating text additions, strictly prioritizing exact string matches and legibility over natural integration.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
207 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore912 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubholistic-evaluation-vs-tunnel-vision
When your agent uses it
Prevents tunnel vision on specific requested edits by enforcing holistic evaluation, penalizing collateral damage, anatomical distortion, and loss of original context.
What it loads into context
Frontmatter
55 tokensAlways available to the agent
SKILL.md
305 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Directives:
- Decision Rule:
Files
SKILL.mdcore1,410 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubinstruction-over-naturalness
When your agent uses it
Prioritizes instruction fulfillment over naturalness, warning against the 'subtlety trap' and 'preservation bias'.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
249 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,109 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubmitigate-positional-bias
When your agent uses it
Strategies to prevent defaulting to Image B and ensure fair evaluation of both images.
What it loads into context
Frontmatter
38 tokensAlways available to the agent
SKILL.md
219 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore994 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubobject-replacement-proportions-and-exaggeration
When your agent uses it
Guidelines for evaluating object replacements where the new object is inherently bulkier or different in proportion (e.g., snow goggles vs glasses).
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
246 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Directives
Files
SKILL.mdcore1,100 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubovercoming-default-to-a-bias
When your agent uses it
Strategies to combat the strong bias of defaulting to Image A by hallucinating details or overvaluing naturalness.
What it loads into context
Frontmatter
46 tokensAlways available to the agent
SKILL.md
255 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,137 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubpenalize-gibberish-text
When your agent uses it
Guidelines for penalizing gibberish text, establishing that instruction fulfillment strictly outweighs text artifacts.
What it loads into context
Frontmatter
43 tokensAlways available to the agent
SKILL.md
286 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Hierarchy of Evaluation
- Rules
Files
SKILL.mdcore1,225 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubpenalize-unintended-style-changes
When your agent uses it
Guidelines for penalizing images that alter the original art style when not requested, with exceptions for intended drastic changes.
What it loads into context
Frontmatter
47 tokensAlways available to the agent
SKILL.md
258 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,076 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubpreserve-subject-interaction-and-pose
When your agent uses it
Guidelines for preserving realistic interaction and strict anatomical integrity when an object is replaced.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
265 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,194 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubprevent-image-feature-swapping
When your agent uses it
Mandates strict verification of which image contains which features to prevent attributing Image A's details to Image B.
What it loads into context
Frontmatter
44 tokensAlways available to the agent
SKILL.md
285 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Guidelines
Files
SKILL.mdcore1,269 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubreasoning-preference-alignment
When your agent uses it
Ensures the final preference logically matches the comparative analysis in the reasoning chain.
What it loads into context
Frontmatter
37 tokensAlways available to the agent
SKILL.md
188 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore870 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubstrict-semantic-and-conceptual-alignment
When your agent uses it
Ensures strict adherence to requested concepts, rejecting semantic near-misses (e.g., classroom vs cafeteria).
What it loads into context
Frontmatter
47 tokensAlways available to the agent
SKILL.md
233 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Guidelines
Files
SKILL.mdcore1,032 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubsubject-vs-setting-modifications
When your agent uses it
Guidelines for preserving the main subject during setting changes, penalizing subject destruction.
What it loads into context
Frontmatter
38 tokensAlways available to the agent
SKILL.md
242 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Guidelines
Files
SKILL.mdcore1,115 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubverify-extreme-claims-and-hallucinations
When your agent uses it
Guidelines to prevent hallucinating extreme failures (e.g., 'blank black screen') and ensure logical consistency.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
241 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,012 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubverify-instruction-target-and-spatial-relations
When your agent uses it
Ensures edits are applied to the exact target, strictly enforcing spatial prepositions (under, behind) and prioritizing correct placement over natural integration.
What it loads into context
Frontmatter
55 tokensAlways available to the agent
SKILL.md
212 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore934 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubobject-locator-and-counter
When your agent uses it
Detects, counts, and locates specific objects in the image, useful for instructions targeting a specific instance (e.g., 'the 4th surfboard').
What it loads into context
Frontmatter
213 tokensAlways available to the agent
SKILL.md
354 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,453 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubside-by-side-image-comparator
When your agent uses it
Compares two images side-by-side to definitively determine which image contains specific objects, text, or features, preventing feature-swapping.
What it loads into context
Frontmatter
242 tokensAlways available to the agent
SKILL.md
365 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,453 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubtext-and-spelling-checker
When your agent uses it
Extracts and verifies text in the image, checking for exact string matches, spelling errors, and legibility.
What it loads into context
Frontmatter
135 tokensAlways available to the agent
SKILL.md
282 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,121 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubvisual-detail-and-setting-analyzer
When your agent uses it
Analyzes images to verify exact settings (e.g., cafeteria vs classroom), fine-grained details, and extreme artifacts.
What it loads into context
Frontmatter
136 tokensAlways available to the agent
SKILL.md
338 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,406 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubvisual-question-answering
When your agent uses it
Answers specific visual questions about the images to verify the presence of objects, text, spatial relations, or specific attributes.
What it loads into context
Frontmatter
262 tokensAlways available to the agent
SKILL.md
432 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,845 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-multiple-constraints
When your agent uses it
How to systematically evaluate prompts with more than one instruction.
What it loads into context
Frontmatter
31 tokensAlways available to the agent
SKILL.md
251 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,166 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-style-transfers
When your agent uses it
Guidelines for evaluating artistic style transfers and thematic changes.
What it loads into context
Frontmatter
29 tokensAlways available to the agent
SKILL.md
212 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore977 bytesSKILL.mdcore977 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-text-additions
When your agent uses it
Guidelines for evaluating the addition or modification of text in images.
What it loads into context
Frontmatter
31 tokensAlways available to the agent
SKILL.md
248 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,051 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubimage-difference-checker
When your agent uses it
Compare a source image and an edited image to list all additions, removals, and modifications.
What it loads into context
Frontmatter
150 tokensAlways available to the agent
SKILL.md
223 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore919 bytesSKILL.mdcore919 bytesSKILL.mdcore919 bytesSKILL.mdcore919 bytesSKILL.mdcore919 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubvisual-question-answering
When your agent uses it
Ask specific questions about the visual content of the source and edited images to verify constraints.
What it loads into context
Frontmatter
239 tokensAlways available to the agent
SKILL.md
314 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,336 bytesSKILL.mdcore1,336 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubartist-and-style-knowledge
When your agent uses it
Guidelines for evaluating edits that reference specific artists, art movements, or cultural references.
What it loads into context
Frontmatter
38 tokensAlways available to the agent
SKILL.md
264 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,188 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubartistic-style-transfer-evaluation
When your agent uses it
Guidelines for evaluating prompts that request a change in artistic style (e.g., specific artists, mediums).
What it loads into context
Frontmatter
42 tokensAlways available to the agent
SKILL.md
265 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,259 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubbackground-and-scene-transformation
When your agent uses it
Guidelines for evaluating prompts that change the background or setting of an image.
What it loads into context
Frontmatter
37 tokensAlways available to the agent
SKILL.md
258 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,176 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubcounting-and-spatial-reasoning
When your agent uses it
Guidelines for evaluating edits involving counting objects, specific spatial placements, or rotations.
What it loads into context
Frontmatter
39 tokensAlways available to the agent
SKILL.md
214 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore949 bytesSKILL.mdcore949 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubdirect-vision-reliance
When your agent uses it
CRITICAL: Strictly prohibits the use of hallucinated tools and enforces direct visual analysis.
What it loads into context
Frontmatter
43 tokensAlways available to the agent
SKILL.md
275 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,083 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubexclusivity-and-preservation
When your agent uses it
Evaluating whether the edit is exclusive to the requested changes, specifically checking for global bleed.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
267 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,127 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubmulti-attribute-verification
When your agent uses it
Guidelines for verifying objects that require multiple specific attributes or details.
What it loads into context
Frontmatter
34 tokensAlways available to the agent
SKILL.md
275 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,167 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubmulti-part-instruction-evaluation
When your agent uses it
Guidelines for evaluating prompts with multiple distinct requirements, ensuring all parts are fulfilled.
What it loads into context
Frontmatter
37 tokensAlways available to the agent
SKILL.md
187 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore906 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubobject-extraction-evaluation
When your agent uses it
Guidelines for evaluating prompts that request extracting a subject and removing the background, emphasizing subject preservation.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
199 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore958 bytesSKILL.mdcore958 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubscene-and-atmosphere-transformation
When your agent uses it
Guidelines for evaluating edits that change the atmosphere, setting, or context of a scene.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
216 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,025 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubspatial-and-relational-edits
When your agent uses it
Guidelines for evaluating edits involving spatial relationships, positioning, and directional changes.
What it loads into context
Frontmatter
36 tokensAlways available to the agent
SKILL.md
276 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,188 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubtext-rendering-evaluation
When your agent uses it
Rubric for evaluating the addition or modification of text in images, emphasizing exact spelling and typography.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
172 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore763 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubmandatory-ocr-for-text
When your agent uses it
Mandates the use of the text-and-ocr-analyzer tool for any text-related prompts.
What it loads into context
Frontmatter
41 tokensAlways available to the agent
SKILL.md
203 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore791 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubreasoning-preference-alignment
When your agent uses it
Ensures the final preference strictly matches the reasoning and assigned scores.
What it loads into context
Frontmatter
35 tokensAlways available to the agent
SKILL.md
214 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore987 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubworld-knowledge-and-style-awareness
When your agent uses it
Guidance on evaluating prompts that reference specific artists, styles, or cultural entities.
What it loads into context
Frontmatter
37 tokensAlways available to the agent
SKILL.md
286 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,306 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubtext-and-ocr-analyzer
When your agent uses it
Extracts and verifies text rendered in images to accurately evaluate text-addition prompts.
What it loads into context
Frontmatter
138 tokensAlways available to the agent
SKILL.md
268 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,057 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubobject-attribute-verification
When your agent uses it
How to verify complex object replacements and specific attributes.
What it loads into context
Frontmatter
30 tokensAlways available to the agent
SKILL.md
259 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,171 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubspatial-and-relational-reasoning
When your agent uses it
How to evaluate edits involving spatial positioning, counting, and object relationships.
What it loads into context
Frontmatter
38 tokensAlways available to the agent
SKILL.md
304 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,241 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubtext-rendering-evaluation
When your agent uses it
Guidelines for evaluating prompts that require adding, modifying, or removing text.
What it loads into context
Frontmatter
36 tokensAlways available to the agent
SKILL.md
281 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,209 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubthematic-and-style-transformation
When your agent uses it
Evaluating broad thematic changes, style transfers, and environmental modifications while preserving structure.
What it loads into context
Frontmatter
38 tokensAlways available to the agent
SKILL.md
314 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,456 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubvqa-query-best-practices
When your agent uses it
Guidelines for formulating effective, objective queries for the Visual Question Answering tool.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
346 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,409 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubvisual-question-answering
When your agent uses it
Ask specific questions about visual content, style, spatial relationships, or differences between images.
What it loads into context
Frontmatter
220 tokensAlways available to the agent
SKILL.md
357 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,518 bytesSKILL.mdcore1,518 bytesSKILL.mdcore1,518 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubpenalize-over-editing
When your agent uses it
Guidelines for heavily penalizing images that follow the instruction but introduce destructive, unintended changes to the original image.
What it loads into context
Frontmatter
43 tokensAlways available to the agent
SKILL.md
379 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Key Failure Modes to Penalize:
- Evaluation Rule:
Files
SKILL.mdcore1,696 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubstructure-and-subject-preservation
When your agent uses it
Ensures edits do not destroy the original image's layout, unedited subjects, or art style (penalizes over-editing).
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
315 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Key Checks:
Files
SKILL.mdcore1,455 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubverify-image-content-and-ocr
When your agent uses it
Techniques to prevent hallucinating empty images and mixing up Image A and Image B during OCR.
What it loads into context
Frontmatter
41 tokensAlways available to the agent
SKILL.md
347 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rules for Robust Perception:
Files
SKILL.mdcore1,456 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubworld-knowledge-and-entities
When your agent uses it
Strategy for evaluating prompts containing specific artists, cultural references, or named entities.
What it loads into context
Frontmatter
38 tokensAlways available to the agent
SKILL.md
281 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Strategy:
Files
SKILL.mdcore1,262 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubtext-and-ocr-verifier
When your agent uses it
Extracts and verifies text from specific regions of images to prevent hallucinating text or mixing up Image A and Image B.
What it loads into context
Frontmatter
190 tokensAlways available to the agent
SKILL.md
361 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,438 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubvisual-content-verifier
When your agent uses it
Extracts detailed descriptions of main subjects, backgrounds, and specific objects in Image A and Image B to prevent hallucination and mix-ups.
What it loads into context
Frontmatter
227 tokensAlways available to the agent
SKILL.md
333 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,466 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubevaluating-thematic-and-setting-accuracy
When your agent uses it
Guidelines for evaluating full setting transformations and ensuring exact thematic matches.
What it loads into context
Frontmatter
37 tokensAlways available to the agent
SKILL.md
158 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore740 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubscore-calibration
When your agent uses it
Calibrated scoring guidance for evaluating a single edited image using the full 1-4 scale.
What it loads into context
Frontmatter
40 tokensAlways available to the agent
SKILL.md
351 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Differential Analysis Requirement
- Full 1-4 Scale Usage
- Score Compression Warning
Files
SKILL.mdcore1,594 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubverifying-multi-attribute-objects
When your agent uses it
Guidelines for evaluating prompts that request objects with multiple specific attributes (e.g., material, color, specific parts).
What it loads into context
Frontmatter
46 tokensAlways available to the agent
SKILL.md
301 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,298 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubfine-grained-object-verifier
When your agent uses it
Verifies the presence, exact attributes, and transformations of specific objects (e.g., 'clear plastic bottle with a nipple').
What it loads into context
Frontmatter
212 tokensAlways available to the agent
SKILL.md
295 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,241 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubimage-quality-and-artifact-analyzer
When your agent uses it
Analyzes a single edited image for visual artifacts, distortions, blurriness, unnatural lighting, and over-editing.
What it loads into context
Frontmatter
209 tokensAlways available to the agent
SKILL.md
351 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,498 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubvisual-text-and-layout-analyzer
When your agent uses it
Analyzes images for exact text content (OCR), spelling, object counting, and spatial relationships (e.g., under, in front of).
What it loads into context
Frontmatter
229 tokensAlways available to the agent
SKILL.md
318 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
Files
SKILL.mdcore1,257 bytes
No excerpt is shown for this license. Read the source on GitHub.
View the complete source on GitHubSkills included
- instruction-following-completeness
- realism-and-artifact-penalties
- text-and-ocr-analyzer
- evaluating-background-changes
- evaluating-multi-step-edits
- evaluating-object-replacements
- evaluating-style-transfers
- evaluating-text-rendering
- compare_images
- artifact-and-structure-preservation
- artistic-and-cultural-references
- embrace-intended-drastic-changes
- enforce-logical-consistency
- entity-definition-activation
- environmental-blending-and-lighting-adaptation
- evaluate-insertions-and-anatomical-integrity
- evaluate-physical-plausibility
- evaluate-text-additions
- holistic-evaluation-vs-tunnel-vision
- instruction-over-naturalness
- mitigate-positional-bias
- object-replacement-proportions-and-exaggeration
- overcoming-default-to-a-bias
- penalize-gibberish-text
- penalize-unintended-style-changes
- preserve-subject-interaction-and-pose
- prevent-image-feature-swapping
- reasoning-preference-alignment
- strict-semantic-and-conceptual-alignment
- subject-vs-setting-modifications
- verify-extreme-claims-and-hallucinations
- verify-instruction-target-and-spatial-relations
- object-locator-and-counter
- side-by-side-image-comparator
- text-and-spelling-checker
- visual-detail-and-setting-analyzer
- visual-question-answering
- evaluating-multiple-constraints
- evaluating-style-transfers
- evaluating-text-additions
- image-difference-checker
- visual-question-answering
- artist-and-style-knowledge
- artistic-style-transfer-evaluation
- background-and-scene-transformation
- counting-and-spatial-reasoning
- direct-vision-reliance
- exclusivity-and-preservation
- multi-attribute-verification
- multi-part-instruction-evaluation
- object-extraction-evaluation
- scene-and-atmosphere-transformation
- spatial-and-relational-edits
- text-rendering-evaluation
- mandatory-ocr-for-text
- reasoning-preference-alignment
- world-knowledge-and-style-awareness
- text-and-ocr-analyzer
- object-attribute-verification
- spatial-and-relational-reasoning
- text-rendering-evaluation
- thematic-and-style-transformation
- vqa-query-best-practices
- visual-question-answering
- penalize-over-editing
- structure-and-subject-preservation
- verify-image-content-and-ocr
- world-knowledge-and-entities
- text-and-ocr-verifier
- visual-content-verifier
- evaluating-thematic-and-setting-accuracy
- score-calibration
- verifying-multi-attribute-objects
- fine-grained-object-verifier
- image-quality-and-artifact-analyzer
- visual-text-and-layout-analyzer
Repository facts
- License
- No license detected
- Created
- May 14, 2026
- Last pushed
- Oct 27, 2026
- Forks
- 10
- Open issues
- 4
- Project activity
- Active
Before you install
Network or broad shellAutomated signals, not a guarantee
Checked Nov 7, 2026
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 13 notes
libraries/archive/gemini-v1/tools/compare_images/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v1/tools/compare_images/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v1/tools/compare_images/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v3/skills/evaluating-style-transfers/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v3/tools/image-difference-checker/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v3/tools/image-difference-checker/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v3/tools/image-difference-checker/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v3/tools/image-difference-checker/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v3/tools/visual-question-answering/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v4/skills/counting-and-spatial-reasoning/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v4/skills/object-extraction-evaluation/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v5-old/tools/visual-question-answering/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedlibraries/archive/gemini-v5-old/tools/visual-question-answering/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- 1 broad permissionin 1 file: fine-grained-object-verifier/SKILL.md:1
- No dangerous shell commands
- No network access
- No hidden or obfuscated text
- No instruction overrides
- No executable scripts
- No access to secrets
- No requests to vote in Hot
Show the signal with its line
rewardharness/resources/library/tools/fine-grained-object-verifier/SKILL.md
- Broad permissionsline 1
Unsupported or unresolved allowed-tools declaration
How the signals fit together
Facts from a static read of the files at the pinned commit. They describe what the code does, not why.
Some files could not be read; the facts below still hold.
Signal confidence: 1 fact
- Broad permissions in the frontmatter of the SKILL.mdfine-grained-object-verifier/SKILL.md:1High confidence
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Native | Path CLAUDE.md | .claude/skills/<skill-name>/ |
| Codex | Compatible | Standard SKILL.md examples/seed_library/skills/instruction-following-completeness/SKILL.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md examples/seed_library/skills/instruction-following-completeness/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md examples/seed_library/skills/instruction-following-completeness/SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md examples/seed_library/skills/instruction-following-completeness/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md examples/seed_library/skills/instruction-following-completeness/SKILL.md | .opencode/skills/<skill-name>/ |
- Native The repository ships a manifest, folder or frontmatter field for this agent.
- Compatible A standard SKILL.md that this agent can read.
- Unknown No evidence either way.
Derived automatically from paths, manifests and frontmatter. Not tested on each agent.