All skills
hoodini avatar

/image-master

@83e2083

Master prompt-engineer for photoreal, artifact-free AI still images on ANY tool (Reve, Midjourney, Flux, GPT-image, Imagen, Nano Banana, Stable Diffusion). Builds prompts that hit National-Geographic-grade realism — true skin/fur texture (no plastic), correct anatomy/hands/faces, physically coherent light/shadows/reflections, clean legible text — while engineering deliberate visual IMPACT (color contrast, composition, awe/adrenaline, a striking point of view). Runs a gated process: lock the point-of-view and build the 8-block Capture Stack BEFORE generating, then run a forensic pre-submit inspection against the known artifact list. Use whenever the goal is a single still image that must look REAL and hold up under close inspection — contest entries, hero shots, product/character/wildlife/architecture/concept art, or any time AI images come out plastic, distorted, or fake. Composes with nano-banana-2 (execution) and director (motion). Hebrew triggers: תמונה, תמונות, פוטוריאליזם, ריאליזם, לייצר תמונה, פרומפט לתמונה, בלי עיוותים, עור פלסטיק, ידיים מעוותות, פרצוף מעוות, חדות, נשיונל ג'יאוגרפיק.

Use this Skill: https://skilld.dev/gh/hoodini/ai-agents-skills/image-master

This session only. Nothing lands on disk.

references12-award-winning-composition.md

≈1.4k tokens on demand. Your agent reads this file only when SKILL.md points to it.

12 — Award-Winning Composition: What Wins + the Stop-Scroll Techniques

Distilled from a cross-study of World Press Photo, Wildlife Photographer of the Year, Sony World Photography, National Geographic, Drone Photo Awards, Ocean / Bird Photographer of the Year.

A. What WINS (the common denominators — engineer all of them)

Story first, then a once-in-a-lifetime moment, then light/craft. Technical perfection is the price of entry, not the win.

  1. ONE identifiable subject carries the whole story (synecdoche) — one boy, one hyena, one horseman; never a crowd.
  2. The decisive / un-reshootable instant — peak action, or the split-second chaos organizes into order.
  3. An emotional or moral hook — stakes (grief, survival, tension, tenderness) felt before analysed.
  4. A perspective never seen before — new altitude, underwater, night, or an alignment nobody else waited for (eclipse, mist, reflection).
  5. Scale & juxtaposition — a tiny living thing against the vast / man-made / empty; opposites in one frame (animal in human ruins).
  6. ONE dramatic light source — blue hour, golden backlight, soft window light, night mist, chiaroscuro. Never flat midday.
  7. Disciplined composition — negative space, leading lines, symmetry, frame-within-a-frame.
  8. The gaze / catchlight / connection — eye contact, or a dignified averted gaze.
  9. Technical perfection — necessary, not sufficient (it only breaks ties).
  10. Restraint over shock — the quiet image that speaks loudly beats the gory one.

B. The stop-scroll techniques (pick 1–2 per frame; copy-paste fragments)

  • Scale juxtaposition: a single tiny figure at the lower edge, dwarfed by an immense [glacier/wave/structure] filling the upper three-quarters, vast negative space, the figure under 3% of the frame, 24mm wide (or 200mm telephoto compressing the giant background behind the tiny subject).
  • Reflection in the EYE (the eye as a mirror) ★: extreme macro of the animal's eye, the whole scene faintly mirrored inside the dark glassy iris — the landscape and the low sun curved in the eye, one sharp round catchlight, the eye tack-sharp and alive, fur soft around it, 100mm macro. Keep the reflection FAINT and plausible (a too-detailed "movie in the iris" reads AI-fake).
  • Water / puddle perfect symmetry: camera centimetres above a still puddle at dusk, perfect mirror symmetry, the scene reflected upside-down as sharp as reality, horizon axis dead-centre, the small puddle making the street look vast.
  • Glass / window reflection: shot through a window slightly off the 90° axis, the subject behind the glass layered with the street reflected on it — two realities overlapping in one frame.
  • Unusual angle: extreme low worm's-eye looking up (subject monumental, heroic) / directly overhead top-down (bodies become graphic shapes/patterns) / Dutch tilt 20° for unease.
  • Layering & depth: three planes — out-of-focus foreground element framing the edge, sharp midground subject, atmospheric haze background, each layer lighter and lower-contrast than the one in front, shot through the foreground.
  • Subject breaks the text plane (3D depth cover) ★: see 09 — state the occlusion explicitly.
  • Negative space / frame-within-frame / leading lines / rule of odds: subject on a third, 70% empty real environment (sky/fog/water); seen through a natural archway framing the figure; a road curving from the corner to a small figure on the horizon; an odd-numbered group, one element breaking the symmetry for tension.
  • Decisive moment: frozen at peak action, every element momentarily aligned, the once-in-a-lifetime split second, fast shutter, candid.
  • Gaze + catchlight: looking directly into the lens, one soft window-shaped catchlight per eye, tack-sharp eyes, instant emotional connection.
  • Selective focus: subject isolated against creamy melted bokeh, f/1.4 85mm, razor-thin focus on the eyes, background lights as round bokeh orbs.
  • Colour device: teal-orange complementary contrast / near-monochrome desaturated scene with ONE saturated accent on a real object / monochrome variations of a single hue.

C. Keep it natural, not gimmicky

  • ONE hero technique per frame; don't stack five.
  • A reflection must MEAN something and read plausible (faint, matched to one real light source, shot slightly off the dead-on mirror axis).
  • Always name a lens + aperture — the single biggest "real photo, not render" lever; it forces physically coherent depth.
  • Match technique to emotion (low angle = power, high = vulnerability, tilt = unease).

D. Portfolio strategy (Range + Vision together)

The contest rewards Range (across style/setting/discipline) AND a distinct point of view. Best play: 10 images that span disciplines (wildlife · environment/scale · portrait/character · architecture/sci-fi · product/still-life · conceptual/surreal · graphic-design depth-cover · decisive-moment action · reflection/symmetry · one emotional gut-punch) BUT share a recognisable signature (e.g. "impossible-real moments, one dramatic light, a soul in the eye") so it reads as ONE artist's voice across many worlds. No two near-duplicates; each earns its slot by a different winning denominator + a different technique. A 10-of-one-theme set scores Vision but forfeits Range — diversify.

Source: SKILL.md on GitHub

No alerts3mo3 checks · Risk SAFE
  • Gen Agent Trust Hub3mo

    The skill is a comprehensive instructional framework for AI image prompt engineering. It consists entirely of Markdown documentation and does not contain any executable code, scripts, or mechanisms for data exfiltration. All external references are informative and point to legitimate AI and photography resources.

  • Socket3mo

    No alerts

  • Snyk3mo

    Risk: LOW · No issues

Signed by skilld at 83e2083. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Steadyupdated 4 months ago

README badge

README badge for hoodini/ai-agents-skills/image-master