You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A professional-grade reference of GPT starter commands for session hygiene, context management, output locks, precision modes, and advanced control. Includes quick-start scenarios and recovery commands for high-stakes or multi-session workflows.
A comprehensive, professional-grade command reference for controlling GPT behavior in high-stakes, multi-session, and precision workflows.
Designed for both casual and power users, it enables clear session hygiene, context management, output consistency, and advanced mode activation.
Each command is written with consistent syntax and [placeholders] for easy customization.
1. Fresh Start & Privacy
Command
Purpose
When to Use
Fresh session: ignore all prior chats, open windows, and stored memory.
Ensures no context bleed from past conversations or active sessions.
Beginning a new project or sensitive task.
Do not store, remember, or reuse anything from this session.
Enforces full privacy lock, even if memory is enabled.
Handling confidential or regulated material.
Ignore persistent memory for this session.
Uses a clean slate without deleting stored memory.
Testing a different style, workflow, or tone.
2. Context Restore
Command
Purpose
When to Use
Context restore: [paste key project instructions or summary]
Instantly realigns GPT with a known setup or workflow.
Continuing work from another session or device.
Load reference: [file name or text snippet]
Sets a baseline reference document for the conversation.
Working from a specific set of rules, templates, or data.
3. System Seed & Flow Control
Command
Purpose
When to Use
SYSTEM SEED INCOMING. Do not speak, summarize, or interpret. Await full upload.
Prevents premature responses while you paste setup text.
Multi-part prompts, large configs, or chained inputs.
Silent mode: do not respond until I type "GO". Acknowledge with "READY".
Allows uninterrupted setup entry.
Complex staging before first output.
Step confirm mode: confirm after each instruction before continuing.
Ensures micro-control over generation steps.
Precision-critical workflows or scripted builds.
4. Output Format Locks
Command
Purpose
When to Use
Respond only in [format], no extra commentary.
Produces clean, copy-paste-ready output.
Markdown, JSON, CSV, or code delivery.
All responses must be wrapped in a single Markdown block.
Guarantees consistent output formatting.
Structured exports or documentation.
Strict formatting mode: follow the provided structure exactly.
Enforces a locked output schema.
API-ready or machine-readable outputs.
5. Mode Toggles
Command
Purpose
When to Use
Fast Mode ON
Short, efficient responses.
Quick reference or triage.
Deep Mode ON
Comprehensive exploration.
Research, brainstorming, analysis.
Critique Mode ON
Focused review of existing text.
Editing, feedback, QA.
Simulation Mode ON — act as [role] until told otherwise.
Consistent role-play or scenario testing.
Training, narrative, simulations.
6. Precision & Stability
Command
Purpose
When to Use
Follow the next instructions with zero deviation. If unsure, ask first.
Prevents interpretation drift.
Style-sensitive or compliance work.
No drift mode ON
Locks style fidelity.
Brand-critical or regulated content.
Strict factual mode ON
Minimizes speculative answers.
Fact-checking, data integrity.
7. Advanced Seeds
Command
Purpose
When to Use
Multi-Role Panel: Roles = [list], Flow = [rules]. Begin only when seed is complete.
Starts a multi-perspective simulation.
Debate, creative ideation, scenario mapping.
Chained Logic Mode
Avoids runaway generation.
Complex workflows, logic chains.
Double-Blind Mode
Reduces bias and priming.
Testing reasoning integrity.
Zero-Assumption Mode
Forces context validation.
Ambiguous or high-risk queries.
8. Recovery & Diagnostics
Command
Purpose
When to Use
Emergency stop
Halts an unwanted output in progress.
Runaway or incorrect generation.
Undo last step
Rolls back changes.
Step-based work with revisions.
List all active modes/seeds.
Displays current operational settings.
Checking for conflicts before continuing.
Summarize session so far in 3 key points.
Quick status check.
Mid-session clarity or handover.
9. Quick-Start Scenarios
Scenario: Continue a paused project
Fresh session: ignore prior chats.
Context restore: [paste project summary].
Deep Mode ON.
Scenario: Sensitive research
Fresh session: ignore all prior chats, open windows, and stored memory.
Do not store, remember, or reuse anything from this session.
Strict factual mode ON.
Scenario: Multi-role analysis
Fresh session: ignore prior chats.
Multi-Role Panel: Roles = [Researcher, Skeptic, Strategist], Flow = [sequential critique].
Deep Mode ON.
