The Callisto GPT Agent is a custom AI assistant built on the ChatGPT platform. It is designed to generate awareness posters, maintain brand consistency, and provide reliable information on security topics. Leveraging a structured knowledge base and detailed agent instructions, Callisto helps users create high-quality, on-brand content efficiently.
- Poster Generation: Produces security-awareness poster content with predefined templates and style guidelines.
- Brand Consistency: Adheres to the Robotti brand color system, poster style guide, and tone primer to ensure all outputs align with company standards.
- Security Expertise: References a curated index of security topics and source integrity guidelines to generate accurate and trustworthy content.
- Flexible Prompts: Accepts user prompts for various security themes / risks and adapts output to HTML for a clean
Print To PDFfunctionality with no margins.
callisto_gpt_agent/
├── README.md # This documentation file
├── agent_instructions/ # Core behavior and prompt instructions
│ ├── callisto_instructions_V2025.05.03.txt
│ └── callisto_instructions_V2025.05.05.md # Latest Version
├── agent_knowledge_documents/ # Structured knowledge for the agent
│ ├── robotti_color_system.md
│ ├── robotti_poster_style.md
│ ├── robotti_poster_template.html
│ ├── robotti_security_topics_index.json
│ └── robotti_tone_primer.md
└── example_awareness_posters/ # Sample outputs for reference / testing
├── HELPFUL ADD-ON, HIDDEN RISK.html # Prompt: Let’s do one on browser extensions that look helpful but expose data.
├── HELPFUL ADD-ON, HIDDEN RISK.pdf
├── TYPING IN TRUST_ THINK BEFORE YOU PROMPT.html # Prompt: Can you make a poster about AI data leaks from public prompt sharing?
├── TYPING IN TRUST_ THINK BEFORE YOU PROMPT.pdf
├── YOUR PERSONAL DEVICE, OUR COMPANY’S RISK.html # Prompt: Write a poster warning about insecure use of SSO on personal devices.
└── YOUR PERSONAL DEVICE, OUR COMPANY’S RISK.pdf
- Setup: Upload the documents contained in the
agent_knowledge_documentsfolder into the ChatGPT GPT Configuration Knowledge section. - Configure: In the ChatGPT GPT configuration screen, copy and paste the latest version of
callisto_instructions_V****.**.**.mdinto theInstructionsfield. - Deploy: Give your iteration of
Callistoa unique name and description. Bonus points if you give us credit.
-
Prompt Examples:
Can you make a poster about AI data leaks from public prompt sharing?Let’s do one on browser extensions that look helpful but expose data.Write a poster warning about insecure use of SSO on personal devices.Can you turn ‘don’t copy sensitive stuff into ChatGPT’ into a security one-pager?Make a print-ready flyer about phishing that baits with anger.Can you explain MFA fatigue as a poster someone would *actually* read?Draft a poster about mobile app permissions in BYOD environments.Help me create a security awareness poster using this quote: "I just wanted it to block ads…"Turn this risk idea into a metaphor-first awareness poster: Teams messages leaking outside the org.What would a Robotti-style one-pager look like for unpatched browser exploits?
-
Expected Output: A complete HTML document with inline CSS matching the Robotti color system and poster style, including title, body content, and a footer with source citations.
- Updating Templates: Modify
robotti_poster_template.htmlto change layout or add new placeholders. - Adding Topics: Edit
robotti_security_topics_index.jsonto include additional security themes / examples for better corporate messaging alignment. - Style Adjustments: Tweak values in
robotti_color_system.mdorrobotti_poster_style.mdto refine typography, spacing, and color usage.
- 2025.05.05: Enhanced Markdown instructions with examples, clarifications, and improved process flow instructions.
- 2025.05.03: Initial instructions (in
.txtformat).
This work is licensed under the Creative Commons Attribution-NoDerivatives 4.0 International License (CC BY-ND 4.0).
© 2025 Robotti Tech Services LLC. Unauthorized modification or use without attribution is prohibited.
Use at your own risk — we assume no liability for misuse, misconfiguration, or any downstream consequences.