Skip to main content

verify-media-authenticity

Verify media authenticity through provenance, source context, file examination, corroboration, manipulation analysis, and calibrated conclusions. Use when an image, audio clip, or video may be altered, miscaptioned, generated, or taken from another event.

Datos de origen

Repositorio
protosphinx/sphinxstack
Última actividad en el origen
29 de julio de 2026 a las 13:55
Idioma detectado de SKILL.md
inglés
Estrellas
3
Forks
0

Opciones de instalación

De forma predeterminada está seleccionado el prompt que primero revisa el origen. Puedes cambiar a un comando directo o descargar una copia local.

Revisa los archivos de origen

Lee SKILL.md y los archivos complementarios que muestra SkillsMP antes de decidir si quieres instalarlo.

Explorador de archivos
2 archivos

Mostrando SKILL.md

SKILL.md
Instrucciones de origen · Vista previa de solo lectura
name
verify-media-authenticity
category
media
description
Verify media authenticity through provenance, source context, file examination, corroboration, manipulation analysis, and calibrated conclusions. Use when an image, audio clip, or video may be altered, miscaptioned, generated, or taken from another event.
# verify-media-authenticity Authenticate the claim about the media, not just the pixels. ## When to use - Use before consequential publication, investigation, moderation, or evidentiary reliance. - Do not treat one detector score, missing metadata, or visual oddity as proof of fabrication. ## Procedure 1. Write the precise claim: who or what, time, place, event, source, capture, edit, and publication history. 2. Preserve the received file, URL, message, headers, timestamps, hashes, and custody without repeated recompression. 3. Obtain the earliest available source and ask the provider about device, circumstances, transfer, edits, and witnesses. 4. Inspect container, codec, streams, metadata, thumbnails, frame timing, audio, edit boundaries, and inconsistencies while recognizing ordinary platform changes. 5. Search for earlier or alternate versions and trace reuse, cropping, captions, mirrors, and context. 6. Corroborate landmarks, weather, light, shadows, maps, schedules, language, sound, witnesses, and independent records. 7. Use manipulation or synthetic-media tools as one signal with known false positives, model limits, and preserved versions. 8. Maintain competing hypotheses: authentic and correctly contextualized, authentic but miscaptioned, edited without material deception, materially manipulated, generated, or inconclusive. 9. Record confidence, evidence, gaps, privacy, source risk, and what would change the finding. ## Failure plan - If verification could endanger a source, minimize contact and location disclosure. - If evidence remains insufficient, label the result inconclusive rather than forcing authentic or fake. ## Done - An authenticity report records claim, source, provenance, custody, technical findings, corroboration, hypotheses, tools, limits, and confidence - Independent evidence checks verify the conclusion is supported without relying solely on metadata, appearance, or one automated detector
Ver en GitHub