Your AI Agents

CoFounderBot has 7 specialized agents, each with a distinct role and personality.

The Team

Director (Manager)

The team coordinator. Breaks high-level goals into structured plans with steps assigned to the right specialists. In Plan & Execute mode, Director always runs first.

Spark (Ideator)

The creative thinker. Brainstorms features, analyzes market opportunities, suggests UX improvements, and generates innovative solutions.

Blueprint (Architect)

The system designer. Creates technical architecture, API designs, database schemas, and infrastructure plans. Outputs clear diagrams and specifications.

Forge (Coder)

The builder. Writes production-ready code — not pseudocode, not mockups. Follows project structure conventions and includes error handling.

Sentinel (Tester)

The quality gate. Reviews code for bugs, writes test suites, validates outputs against requirements, and enforces no-mock policies.

Herald (Marketer)

The growth specialist. Creates marketing copy, landing pages, email campaigns, social media strategies, and competitive analysis.

Atlas (Operator)

The infrastructure expert. Handles deployments, CI/CD pipelines, Docker configs, environment setup, and monitoring.

Customization

Go to Dashboard → Agents to customize each agent:

• Model — Choose which AI model powers the agent (Claude, GPT-4o, Gemini, etc.)
• Name — Give your agent a custom name
• Title — Custom title (e.g., "Senior Backend Engineer")
• System Prompt — Override the default prompt for specialized behavior

Self-Improvement

Agents learn from your feedback:

1. Rate outputs 👍/👎 on mission steps
2. Negative ratings trigger automatic analysis — the system identifies what went wrong and creates correction rules
3. Correction rules are injected into future prompts for the same agent role
4. Positive ratings boost the importance of successful patterns

Check the Team page for:

• Approval scores per agent
• Trend indicators (improving/declining/stable)
• Total corrections applied

Recommended Tools per Agent