How to spawn projects from a single file using AI

The power of the "Seed" concept lies in its "AI Autonomy Directive", which instructs the AI to:

• Orient Itself: The AI first locates and parses project status files like DEVELOPMENT_STATUS.json and PROMPT_TRACKING.json to understand the current project status, the last_completed_prompt_id, and the next_planned_prompt_id. If these files don't exist, the AI recognizes it's a fresh start. It then thoroughly reviews the "Active Instruction Document" (the Seed file) and its parsed derivatives to grasp project goals, architecture, module responsibilities, and the specifics of the next task.

• Acknowledge Understanding: The AI confirms its understanding of the project's current status and the immediate next task.

• Check Assets: It verifies available assets against a user_assets_manifest.json.

• Execute Autonomously:

• If DEVELOPMENT_STATUS.json indicates a next_planned_prompt_id, the AI immediately begins that task.

• For new projects without existing status files, the AI identifies the initial setup task (e.g., TASK-SYS-001) from the Seed file and autonomously develops and proposes the necessary script (e.g., project_init.py) without requiring further permission.

• For all subsequent tasks, the AI executes the next_planned_prompt_id defined in DEVELOPMENT_STATUS.json.

• Throughout execution, the AI uses the Seed file and its derivatives as the primary guide, adhering to a defined "Autonomous Execution Policy". The AI is explicitly instructed not to ask for permission for routine or well-defined steps, only pausing for genuine ambiguities, missing information, or critical decision points not covered by the plan.

The Seed concept incorporates "Core Behavioral Rules" to govern the AI's actions:

• Autonomous Progression: The AI progresses autonomously through tasks and phases outlined in tracking files.

• Human Intervention Policy: The AI only pauses or requests human intervention if explicitly required, if a critical ambiguity prevents progress, or if a tool repeatedly fails.

• Tool Usage: It utilizes defined helper CLI tools, with direct file manipulation as a fallback if tools fail.

• Tracking File Management: The AI proactively manages PROMPT_TRACKING.json, DEVELOPMENT_STATUS.json, and HUMAN_ACTION_ITEMS.json.

• Parsed File Preference: After initial setup, the AI prefers using parsed output files for task details.

A crucial element for handling complex projects is the "AI-Driven Sub-Task Generation Protocol". When the AI identifies an active prompt as too complex, it:

• Consults the Plan: Reviews the Seed file and core principles to ensure alignment.

• Defines Sub-Prompts: Creates 2-5 smaller, actionable sub-prompts, each with a clear description and Definition of Done (DoD) criteria. This process can be recursive.

• Assigns IDs: Assigns unique IDs to each sub-prompt.

• Logs Sub-Prompts: Adds these sub-prompts to PROMPT_TRACKING.json.

• Updates Parent: Changes the status of the original complex prompt to "Decomposed" or "Blocked by Sub-tasks".

• Proceeds Sequentially: Makes the first generated sub-prompt its active task.