Text Watermark Fountain

A specialized skill for embedding and extracting resilient watermarks in text by manipulating sentence lengths and using Fountain Codes. Use this when the user wants to add a hidden, robust watermark to text or verify an existing one.

260 stars 30 forksAI, LLMs & Agents

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add cafe3310/public-agent-skills --skill text-watermark-fountain
Manual install
git clone --depth 1 https://github.com/cafe3310/public-agent-skills.git /tmp/public-agent-skills
mkdir -p ~/.claude/skills
cp -r /tmp/public-agent-skills/skills_parked/text-watermark-fountain ~/.claude/skills/text-watermark-fountain

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Text Watermark Fountain (Robust Sync-Frame Version)

This skill enables the Agent to embed a string watermark into a text such that it can be recovered even if the text is partially modified, segments are deleted, or new sentences are inserted. It uses a custom Luby Transform (LT) Fountain Code combined with **Sync Frames** to map the watermark into a sequence of target lengths.

## How it works (Robustness Mechanism)

1.  **Sync Markers**: The encoding script periodically inserts a unique length pattern `[19, 4, 19]` (Sync Marker) followed by a **Frame ID**.
2.  **Self-Synchronization**: The decoder searches the entire text for these markers using a sliding window. Even if middle segments are removed, the decoder can resynchronize using the next Sync Marker and know exactly which symbols it is looking at.
3.  **Redundancy**: By repeating these frames throughout a long text, the watermark becomes extremely difficult to destroy.

## Workflow: Embedding a Watermark

When a user asks to embed a watermark (e.g., "name_1") into a text:

1.  **Generate Length Sequence**:
    - Run the encoding script:
      ```bash
      python3 scripts/encode.py --mark "name_1" --count [TOTAL_DATA_SYMBOLS]
      ```
    - Note the `Lengths` output. It will contain periodic `19 4 19 [ID]` headers.

2.  **Precise Text Fine-tuning (Batch Processing)**:
    To ensure 100% accuracy, you MUST use a multi-stage approach:
Read the full skill on GitHub

Details

Skill name
text-watermark-fountain
Path
skills_parked/text-watermark-fountain/SKILL.md
License
Apache-2.0
Repo updated
15 November 2026

FAQ

What does the Text Watermark Fountain skill do?

A specialized skill for embedding and extracting resilient watermarks in text by manipulating sentence lengths and using Fountain Codes. Use this when the user wants to add a hidden, robust watermark to text or verify an existing one.

How do I install the Text Watermark Fountain skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add cafe3310/public-agent-skills --skill text-watermark-fountain. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Text Watermark Fountain popular?

It lives in cafe3310/public-agent-skills, which has 260 GitHub stars and 30 forks.

Related skills

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.