C2pa Metadata logo

C2pa Metadata

Community
indranilbanerjee
c2pa-metadata

Embed C2PA provenance in AI-generated images, video or PDF by script. "add content credentials to this asset"

Overview

Publisherindranilbanerjee
Repositorydigital-marketing-pro
Skill namec2pa-metadata
Stars
866
Forks
138
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by indranilbanerjee on GitHub. Read the source before you install it.

Installation

Install the C2pa Metadata AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/indranilbanerjee/digital-marketing-pro/tree/main/skills/c2pa-metadata
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/indranilbanerjee/digital-marketing-pro.git /tmp/digital-marketing-pro
mkdir -p .claude/skills
cp -r /tmp/digital-marketing-pro/skills/c2pa-metadata .claude/skills/c2pa-metadata
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable C2pa Metadata in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads C2pa Metadata on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so C2pa Metadata is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

/digital-marketing-pro:c2pa-metadata — Embed Content Authenticity Provenance

Purpose

Wraps scripts/embed-c2pa.py to add a C2PA (Coalition for Content Provenance and Authenticity) manifest to any AI-generated marketing asset. The manifest carries a machine-readable provenance trail (who generated it, what generator was used, what prompt produced it, when it was reviewed) plus a visible AI-generation claim in the IPTC digital-source-type vocabulary.

This is the technical mechanism brands use to comply with:

  • EU AI Act Article 50 (applicable 2 August 2026) — generative-AI marketing content must be marked in a machine-readable format using open, interoperable standards. C2PA is the emerging backbone. Penalty for non-compliance: up to €15 million or 3% global annual turnover.
  • NY synthetic-performer disclosure law (effective June 2026) — $1K–$5K per violation, $10K repeat; applies to synthetic influencers and AI-generated endorsements.
  • FTC May 2026 endorsement guidance — covers AI testimonials and synthetic creator content.
  • Australia Online Safety Act / UK Online Safety Act — emerging deepfake disclosure requirements.

The resulting asset can be inspected by any C2PA-aware viewer (Adobe Photoshop, Lightroom, Truepic, contentcredentials.org/verify).

C2PA spec versions to be aware of (June 2026)

  • Content Credentials 2.3 (released 9 February 2026 — launch post) added format support for: live video (broadcast/streaming), plain text documents, OGG Vorbis audio, large AVI video files, and EXIF Original Preservation Images. If a brand is signing live-stream video or text-based assets for the first time, 2.3 is the floor version to target.
  • C2PA Spec 2.4 (April 2026 — spec.c2pa.org/specifications/specifications/2.4) introduces the AI Disclosure Assertion (c2pa.ai-disclosure) for machine-readable AI transparency info — this is the assertion the EU AI Act Article 50 deployer pathway will rely on. The final Code of Practice on Transparency of AI-Generated Content (published 10 June 2026) references C2PA-style assertions as the canonical machine-readable marking mechanism for both providers and deployers. See skills/context-engine/eu-code-of-practice.md for the full Article 50 context.
  • The C2PA Trust List is now handled via the public C2PA Conformance Program (any CA meeting the Certificate Policy can join). Production signing certificates should come from a Conformance-Program-listed CA, not an ad-hoc cert.

For DMP outputs: embed-c2pa.py now supports --ai-disclosure. Pass it to embed the C2PA 2.4 c2pa.ai-disclosure assertion alongside the existing IPTC digital-source-type claim. The combination gives you both human-readable (IPTC) and machine-readable (c2pa.ai-disclosure) EU AI Act Article 50 signaling — this is the deployer-side machine-readable pathway the final Code of Practice (10 June 2026) points to as the canonical marking mechanism. See skills/context-engine/eu-code-of-practice.md for the full Article 50 context.

When to invoke

  • Right after any AI image / video / audio generation step in the engagement workflow (Part 11 — AI Creative Instructions output)
  • Before handing a generated asset to the design team for review
  • As a pre-publish gate in /digital-marketing-pro:check for EU-targeted assets
  • Bulk-applying to a backlog of AI-generated assets before EU AI Act enforcement on 2 Aug 2026

Quick examples

bash
# Single asset — image generated by Vertex AI / Nano Banana Pro
/digital-marketing-pro:c2pa-metadata \
    --input assets/q3-launch-hero.png \
    --output assets/signed/q3-launch-hero.png \
    --brand "Acme Corp" \
    --generator "Vertex AI / Nano Banana Pro" \
    --ai-claim ai-generated-content \
    --prompt "minimalist product hero shot, soft natural lighting"

# Video with human review tracked
/digital-marketing-pro:c2pa-metadata \
    --input campaigns/launch-video-v3.mp4 \
    --output campaigns/signed/launch-video-v3.mp4 \
    --brand "Acme Corp" \
    --generator "Runway Gen-4" \
    --ai-claim ai-generated-content \
    --reviewer "Jane Smith"

# EU-targeted asset — add the machine-readable Article 50 AI-disclosure assertion (C2PA 2.4)
/digital-marketing-pro:c2pa-metadata \
    --input assets/q3-launch-hero.png \
    --output assets/signed/q3-launch-hero.png \
    --brand "Acme Corp" \
    --generator "Vertex AI / Nano Banana Pro" \
    --ai-claim ai-generated-content \
    --ai-disclosure \
    --prompt "minimalist product hero shot, soft natural lighting"

# Human-created image with AI-assisted edits
/digital-marketing-pro:c2pa-metadata \
    --input assets/founder-headshot-edited.jpg \
    --output assets/signed/founder-headshot-edited.jpg \
    --brand "Acme Corp" \
    --generator "Adobe Generative Fill" \
    --ai-claim ai-assisted-edits

# Production sign with a real C2PA signing certificate
/digital-marketing-pro:c2pa-metadata \
    --input assets/q3-launch-hero.png \
    --output assets/signed/q3-launch-hero.png \
    --brand "Acme Corp" \
    --generator "Vertex AI / Nano Banana Pro" \
    --ai-claim ai-generated-content \
    --signing-cert /secure/c2pa-prod-cert.pem \
    --signing-key /secure/c2pa-prod-key.pem

AI claim values (IPTC digital source type)

ValueWhen to useMaps to IPTC URI
ai-generated-contentAsset fully generated by AIalgorithmicMedia
ai-assisted-editsHuman-created + AI-edited (e.g. Generative Fill)compositeWithTrainedAlgorithmicMedia
ai-no-substantive-changesAI used (e.g. upscaling) but no semantic changeminorHumanEdits

The IPTC vocabulary is what EU AI Act regulators reference — using these values rather than ad-hoc strings makes the asset interoperable with the Article 50 enforcement tooling.

Supported asset formats

.png · .jpg/.jpeg · .webp · .gif · .tiff · .mp4 · .mov · .webm · .mp3 · .wav · .pdf

Signing certificate

Production C2PA signatures require a certificate from a CAI-recognized signing authority. The script will use one if you pass --signing-cert and --signing-key. If you omit them, the script generates a self-signed 90-day dev certificate for development testing only — a self-signed asset will verify as "signature present but signer not in trust list" at contentcredentials.org/verify.

For production deployment:

  1. Obtain a C2PA-compatible signing certificate from a CAI-recognized authority (Adobe, Truepic, Numbers Protocol, Microsoft Azure Confidential Ledger).
  2. Store the cert + key securely (do NOT commit to git; use an environment-variable path or secret store).
  3. Pass --signing-cert and --signing-key on every production invocation.

Reference: opensource.contentauthenticity.org/docs/manifest/signing-manifests/

Python dependencies

  • c2pa-python==0.38.0: required for signing and for --verify
  • cryptography==46.0.3: only needed for the dev self-signed cert path

The script never installs anything. If either package is missing it prints the exact python -m pip install "…" command and exits with code 2; show that command to the user and let them decide whether to run it. Do not run pip on their behalf.

Each signature requests a timestamp from http://timestamp.digicert.com (a hash of the claim, not the asset). The dev key, when used, is deleted as soon as the script exits.

Output

The script prints a JSON status report to stdout:

json
{
  "status": "success",
  "input": "assets/q3-launch-hero.png",
  "output": "assets/signed/q3-launch-hero.png",
  "size_bytes": 482371,
  "brand": "Acme Corp",
  "generator": "Vertex AI / Nano Banana Pro",
  "ai_claim": "ai-generated-content",
  "created": "2026-05-16T10:30:00+00:00",
  "manifest_assertions": ["c2pa.actions", "stds.schema-org.CreativeWork"],
  "using_dev_cert": false,
  "verify_url": "https://contentcredentials.org/verify"
}

Integration with the engagement workflow

In a full 12-part engagement, this skill plugs in at Part 11 — AI Creative Instructions output. After a creative brief is rendered as an actual asset (by your creative tooling or a manual creative process), the resulting file passes through c2pa-metadata before being checked in to engagements/<slug>/11-creative-briefs/signed/.

The /digital-marketing-pro:check pre-publish gate should also verify that all AI-generated assets in an EU-targeted campaign carry a C2PA manifest. v3.4 adds this verification to the EU jurisdiction rule pack in skills/context-engine/compliance-rules.md.

Related

Frequently asked questions

What does the C2pa Metadata AI skill do?

Embed C2PA provenance in AI-generated images, video or PDF by script. "add content credentials to this asset"

Why use C2pa Metadata on TypingMind?

Because you install it once and use it with any model. C2pa Metadata is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install C2pa Metadata in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/indranilbanerjee/digital-marketing-pro/tree/main/skills/c2pa-metadata. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use C2pa Metadata?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with C2pa Metadata?

As many as you like. As long as a model supports skills, you can use C2pa Metadata with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the C2pa Metadata AI skill free?

Yes. It is published on GitHub by indranilbanerjee under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇