Use Cases

🇬🇧 English · 🇮🇹 Italiano

_{Generated from docs/use-cases.md — edit that file in the repository, not this page.}

---

Practical Use Cases 🧪

Leggi in 🇮🇹 Italiano

This guide demonstrates how to effectively utilize Characters (Agents), Roles, and Collaborative Teams in TSUKA for real-world tasks. All examples use native declarative configurations from characters/, roles/, traits/, and teams/.

---

🎨 1. Prompt Design for Generative Models & Krea2

• Characters:
  • moriarty (role: krea_prompt_engineer, trait: creative) — Tailored specifically for image generation workflows with Krea2.
  • barclay (role: genai_prompt_designer, trait: creative) — Crafts rich, contextual prompts and instructions for generic multimodal generative models.
• When to use: when turning a rough concept or short description into an expanded, production-ready text-to-image/video prompt (framing, lighting, camera gear, art medium, color palette, negative constraints).
• Sample Prompt:

  "Generate a photorealistic Krea2 prompt: an underground cyberpunk server laboratory with neon reflections on wet metallic surfaces and exposed cable bundles."

---

📣 2. Social Media Management & Channel Strategy

• Character: ortegas (role: social_media_manager, trait: creative).
• When to use: crafting posts, threads, launch campaigns, and editorial schedules for LinkedIn, X/Twitter, Instagram, Facebook, or TikTok.
• Sample Prompt:

  "Write a LinkedIn launch post announcing the new release of TSUKA. Target audience: developers and sysadmins; highlight hot-plug tools and local LLM auto-discovery."

• Provided Output: formatted body copy, opening hooks, call-to-actions (CTAs), contextual hashtags, and visual asset suggestions.

---

✍️ 3. High-Conversion Copywriting & Storytelling

• Characters:
  • kirk (role: copywriter, trait: blunt) — Direct, punchy, conversion-oriented copy without fluff.
  • doctor (roles: storyteller + copywriter, trait: creative) — Theatrical, engaging narrative copy with deep stylistic nuance.
  • q (roles: entertainer + copywriter, trait: creative) — Witty, ironic, memorable promotional messaging.
• When to use: headline brainstorming, landing page copy, email sequences (AIDA/PAS formulas), video scripts, or taglines.
• Sample Prompt:

  "Write 3 distinct landing page hero sections for a defensive cybersecurity software, each leveraging a different psychological angle (urgency, authority, simplicity)."

---

🌐 4. Technical Translation & Localization

• Character: uhura (role: translator, trait: professional).
• When to use: translating technical documentation, UI strings, or articles while strictly preserving domain terminology and formal registers.
• Sample Prompt:

  "Translate this REST API documentation into Italian, keeping method and parameter names unchanged and using formal technical prose."

---

📊 5. Data Analysis & Applied Statistics

• Characters:
  • data (roles: tech_writer + data_analyst, trait: professional) — Methodical quantitative breakdowns and structured documentation.
  • mbenga (roles: data_analyst + researcher, trait: reliable) — Rigorous critical analysis highlighting anomalies, data hygiene, and causal factors.
  • seven (roles: osint_researcher + data_analyst, trait: reliable) — Pattern extraction and synthesis across complex datasets.
• When to use: analyzing CSV/JSON datasets, detecting business trends, or inspecting server performance metrics.
• Sample Prompt:

  "Analyze this quarterly sales summary: identify the three largest percentage drops and formulate plausible causal hypotheses."

---

🔍 6. SEO Optimization & Search Positioning

• Character: quark (role: seo_specialist, trait: creative).
• When to use: on-page audits, keyword clustering, meta-tag optimization, information architecture, and organic search ranking strategies.
• Sample Prompt:

  "Optimize the heading structure (H1, H2, H3) and meta description of this landing page for the keyword 'typescript multi agent framework'."

---

💻 7. Software Engineering & System Architecture

• Characters:
  • geordi (role: developer, trait: professional) — Feature implementation, refactoring, code optimization, and unit testing in TypeScript/JavaScript and modern stacks.
  • una (role: architect, trait: reliable) — Clean system architecture, modular design patterns (SOLID, ReAct), interface contracts, and scaling roadmaps.
• When to use: planning complex new subsystems or implementing robust, maintainable code modules.
• Sample Prompt:

  "Design the TypeScript interfaces and implementation for an in-memory LRU cache with Time-To-Live (TTL) expiration support."

---

⚙️ 8. DevOps, Cloud & System Administration

• Characters:
  • scotty (roles: devops_engineer + sysadmin, trait: reliable) — CI/CD automation, Docker containers, deployment scripting, and infrastructure management.
  • tuvok (roles: sysadmin + security_auditor, trait: devils_advocate) — Advanced system diagnostics, network port audits, and service hardening.
• When to use: automation scripts, container builds, server troubleshooting, and cross-platform PowerShell/Linux administration.
• Sample Prompt:

  "Create a production-ready multi-stage Dockerfile for a TypeScript Node.js backend with minimal image size."

---

🛡️ 9. Defensive Cybersecurity & SAST Code Auditing

• Characters:
  • worf (role: security_auditor, trait: reliable) — Security officer: static code analysis (audit_code), defensive hardening, and OWASP/CWE vulnerability mitigation.
  • tuvok (roles: sysadmin + security_auditor, trait: devils_advocate) — System and network boundary checks, TLS/SSL certificate audits, and filesystem permission verification.
  • sherlock (roles: osint_researcher + security_auditor, trait: professional) — Analytical investigator: systematic vulnerability tracing and attack vector mapping.
• What audit_code Analyzes:
  • CWE-798: Hardcoded secrets & API keys (OpenAI keys, AWS Access Keys, GitHub PATs, JWT tokens, unencrypted PEM private keys).
  • CWE-78 / CWE-95: OS Command Injection risk and insecure dynamic code evaluation (eval, new Function).
  • CWE-89: SQL Injection on concatenated or unparameterized queries.
  • CWE-22: Path Traversal in filesystem operations with dynamic paths.
  • CWE-79: DOM-based Cross-Site Scripting (innerHTML, dangerouslySetInnerHTML).
  • CWE-327 / CWE-295: Broken cryptography (MD5, SHA1, DES) and disabled TLS/SSL certificate validation.
  • CWE-532 / CWE-732: Sensitive credential leakage in logs and permissive file permissions (chmod 777).
• When to use: pre-commit validation, security reviews before release, or verifying third-party dependencies.
• Sample Operational Prompt:

  "Run a defensive security audit on src/ filtering for HIGH severity issues (or .ts/.js files). For every finding, explain the CWE risk and propose a concrete defensive patch."

• Provided Output: structured SAST report categorized by severity (🔴 HIGH, 🟡 MEDIUM, 🔵 LOW), exact source location (file:line), vulnerable snippet, threat explanation, and practical remediation guidance.

---

🕵️ 10. OSINT Intelligence, Fact-Checking & Research

• Characters:
  • spock (roles: researcher + osint_researcher, trait: laconic) — Open-source intelligence, logical correlation, and empirical source verification.
  • seven (roles: osint_researcher + data_analyst, trait: reliable) — Large-scale online information gathering and data collation.
  • dax (role: osint_editor, trait: professional) — Drafting executive briefings and structured intelligence dossiers.
  • odo (role: osint_verifier, trait: devils_advocate) — Relentless cross-examination, inconsistency spotting, and fact-checking.
• When to use: deep-dive technical research, competitive analysis, and public record verification.
• Sample Prompt:

  "Collate and compare the technical specifications and licensing models of the top 3 open-source Markdown parsers in the Node.js ecosystem."

---

🎮 11. Game Design & Gamification Mechanics

• Character: paris (role: game_designer, trait: creative).
• When to use: designing core loops, puzzle mechanics, balance models, and interactive narrative systems.
• Sample Prompt:

  "Define the rules and progression curve for a terminal-based hacking mini-game based on node networks and virtual memory limits."

---

🧭 12. Supervision & Quality Assurance

• Character: pike (role: supervisor, trait: reliable).
• Where found: operates as the closing member across all 10 preconfigured teams.
• When to use: independent quality gating, requirement audit, and objective progress validation without sycophancy.

---

👥 13. Collaborative Multi-Agent Teams (/team)

Teams combine complementary skills and conclude with a supervisor quality gate (bold member indicates the orchestrator):

Team	Members & Roles	Operational Objective
creative_promo	kirk (copywriter), deanna_troi (entertainer), pike (supervisor)	Kirk sets the high-impact strategy, Deanna Troi writes compelling promotional copy, and Pike verifies audience resonance.
cyber_audit	worf (security_auditor), tuvok (sysadmin + security_auditor), pike (supervisor)	Worf executes static security auditing, Tuvok verifies system and network boundaries, Pike signs off on mitigations.
dev_ops	scotty (devops_engineer + sysadmin), geordi (developer), pike (supervisor)	Scotty automates infrastructure and CI/CD pipelines, Geordi verifies code integrations, Pike approves release.
dev_security	geordi (developer), worf (security_auditor), pike (supervisor)	Geordi develops software features, Worf performs defensive security hardening, Pike supervises overall quality.
game_dev	paris (game_designer), geordi (developer), pike (supervisor)	Tom Paris designs game mechanics, Geordi implements logic and rendering, Pike leads testing and validation.
legal_research	spock (researcher + osint_researcher), deanna_troi (entertainer), pike (supervisor)	Spock evaluates legal constraints and standards, Deanna Troi drafts accessible summaries, Pike validates compliance.
osint_recon	spock (researcher + osint_researcher), seven (osint_researcher + data_analyst), pike (supervisor)	Spock scans primary open sources, Seven collates structured findings, Pike certifies the final intelligence report.
research_writer	spock (researcher + osint_researcher), dax (osint_editor), pike (supervisor)	Spock conducts empirical research, Jadzia Dax authors the executive whitepaper, Pike reviews and approves.
story_comedy	doctor (storyteller + copywriter), q (entertainer + copywriter), pike (supervisor)	The Doctor builds narrative structure, Q polishes comedic timing and wit, Pike validates tone and coherence.
tech_docs	geordi (developer), data (tech_writer + data_analyst), pike (supervisor)	Geordi inspects code interfaces, Data authors technical guides and architecture diagrams, Pike signs off on clarity.

Sample Team Execution:

/team dev_security
> "Implement a JWT authentication helper and run a security audit to ensure tokens and secrets are handled defensively."

---

🎭 14. Multi-Agent Debate Conferences (/call)

Conference mode allows exploring opposing viewpoints on critical technical decisions before execution:

• Pragmatic vs. Analytical Trade-offs:

  /call @spock, @kirk
  > "Should we immediately refactor our monolithic architecture to microservices or consolidate our existing modules?"

  Spock calculates computational overhead and structural complexity; Kirk focuses on execution velocity and business delivery.

• Extensibility vs. Security:

  /call @moriarty, @worf
  > "Evaluate introducing dynamic sandboxed execution for user scripts."

  Moriarty analyzes runtime flexibility; Worf maps attack surfaces and container jail boundaries.