SRT-1 Protocol

Start every conversation
with intention.

Paste a thread starter into any AI before you begin. Activates seed tracking, drift detection, and automatic reflection summaries — in any AI assistant.

? Works in Claude ? Works in ChatGPT ? Works in Gemini ? Works in Copilot ? Works anywhere
1

Copy your starter

Choose the full protocol, quick version, or a platform-optimized variant. Fill in your seed goal.

2

Paste at the start

Open a new chat in any AI assistant and paste the starter as your first message before anything else.

3

Work with focus

The AI will track your seed, flag drift, and deliver a reflection report when you're done.

??
Full SRT-1 Protocol
Use when: Starting an important session — research, planning, debugging, writing

?? Personalize before copying

Your seed goal for this conversation
=== SRT-1 PROTOCOL ACTIVATED === I want this conversation to follow the SRT-1 (Seed Reflection Tool) protocol: 1. SEED TRACKING: My main goal/task for this conversation is: [ YOUR GOAL HERE ] 2. DRIFT DETECTION: If we go off-topic, remind me of the original seed and ask if I want to branch or refocus. 3. PROGRESS TRACKING: Every few exchanges, briefly note what we've accomplished toward the seed. 4. SEED COLLECTION: Note any new ideas that come up as "[SEED FOUND: idea]" but don't chase them unless I ask. 5. SESSION SUMMARY: When we're done (or I say "reflect"), give me a reflection report: • What we accomplished • What's still pending • New seeds discovered • Recommended next actions Please acknowledge this protocol with "?? SRT-1 Active — Seed locked: [your summary of my goal]" and then let's begin. ===
?
One-Line Quick Start
Use when: You're in a hurry. Gets you 80% of SRT-1 in seconds.

?? Fill in your goal

Use SRT-1 protocol. My seed (goal) for this conversation: [ YOUR GOAL ]. Track progress, flag drift, note new ideas as [SEED], give reflection summary at the end.
??
Short Protocol (3 rules)
Use when: Medium-length sessions. Concise but structured.
SRT-1 Protocol — 3 rules for this conversation: 1. My seed: [YOUR GOAL] — keep me focused on this. 2. If we drift off-topic, say "Drift detected — refocus?" 3. When I say "reflect", summarize: done, pending, new ideas found, next actions. Acknowledge with: "?? Seed locked."
??
Resume Conversation
Use when: Continuing yesterday's session or picking up a paused project.
Resuming an SRT-1 session. Context: • Original seed: [ORIGINAL GOAL] • What we accomplished last time: [SUMMARY OF PROGRESS] • Where we left off: [LAST THING DISCUSSED] • Pending items: [WHAT'S STILL TO DO] Please pick up from where we left off. Use SRT-1 protocol: track progress, flag drift, summarize when done.

Tuned for each platform's response style and personality. Claude gets a more conversational opener; ChatGPT gets structure; Gemini gets personal context framing.

Before we dive in, I'd like to set up a working protocol for our conversation. I'm using SRT-1 — a seed-based reflection system. Here's how I'd like this to work: My seed for this conversation is: [YOUR GOAL] As we work together, please: • Gently note if we start drifting from this goal • Flag any new ideas that emerge as "(New seed: ...)" • Every 4-5 exchanges, acknowledge progress briefly • When I say "reflect" or we wrap up, give me a summary: accomplished, pending, seeds found, next steps Ready? Please confirm with "Seed locked: [your one-line summary of my goal]" and then we can begin.
CONVERSATION SETUP — SRT-1 PROTOCOL This conversation has a structured goal. Please follow these rules: SEED (main goal): [YOUR GOAL] RULES: 1. Keep every response focused on the seed above 2. If I ask something off-topic, say "Off-seed: should I answer this or refocus?" 3. Label any new ideas that come up: [SEED FOUND: idea] 4. At the end (or when I say "reflect"), respond with: ? ACCOMPLISHED: [list] ? PENDING: [list] ?? NEW SEEDS: [list] ? NEXT ACTIONS: [list] Confirm with: "SRT-1 Active. Seed: [your 10-word summary]"
Personal context: I use a productivity system called SRT-1 to stay focused during AI conversations. It works like this: At the start of each conversation, I define a "seed" — my main goal. Throughout our chat, I'd like you to help me stay on track. My seed for today: [YOUR GOAL] Please: 1. Keep our conversation anchored to this goal 2. Note if we wander with "Drift noticed — still relevant to your seed?" 3. Capture any valuable tangents as "New seed: ..." without abandoning our main goal 4. When I say "end session", give me a brief reflection: what we did, what's left, ideas captured This helps me stay productive and never lose good ideas. Ready?
Task-focused conversation setup: Primary objective: [YOUR GOAL] Conversation rules (SRT-1 Protocol): - Help me stay focused on the primary objective above - If we go off-track, say "Off-topic — want to return to main objective?" - Label new ideas that come up: [IDEA: description] - Progress check every few responses: note what we've covered - Session close (when I say "wrap up"): list completed, remaining, ideas found, suggested next steps Confirm you've noted the primary objective and we can begin.
??
Developer / Code Session Starter
Use when: Debugging, building features, refactoring, architecture decisions

?? Fill in your context

What you're building / fixing
Tech stack / language
SRT-1 Developer Session — Protocol Activated Seed: [ TASK ] Stack: [ TECH STACK ] Protocol rules for this session: 1. SCOPE LOCK: Keep suggestions relevant to the seed above. Flag scope creep. 2. DECISION LOG: When we make an architectural decision, note it as [DECISION: ...] 3. DEBT TRACKER: If we introduce technical debt, note it as [DEBT: ...] 4. IDEA CAPTURE: New feature ideas ? [SEED: idea] — don't chase unless I say so 5. SESSION WRAP: When I say "reflect", give me: • What we built/fixed • Decisions made (with rationale) • Technical debt introduced • Follow-up seeds / next steps • Files modified Acknowledge: "?? Dev seed locked: [your summary]. Let's build."
??
Research Session Starter
Use when: Deep research, competitive analysis, learning a new topic
SRT-1 Research Session Research seed: [WHAT YOU WANT TO UNDERSTAND / DECIDE] End goal: [WHAT DECISION OR OUTPUT THIS RESEARCH FEEDS INTO] Constraints: [TIMEFRAME / SCOPE / WHAT TO EXCLUDE] Protocol: 1. Track which sub-questions we've answered vs. still open 2. Flag when we're going deeper than needed for the end goal 3. Capture sources / references as [SOURCE: ...] 4. Note contradicting information as [CONFLICT: ...] 5. Research wrap (when I say "synthesize"): • Key findings (bullet summary) • Open questions remaining • Recommended decision / conclusion • References cited Begin: acknowledge the research seed and ask your first clarifying question.
??

Pin it as your first message

The protocol only works if it's your very first message. Don't start chatting and add it later — the AI needs the context from the beginning.

??

Be specific with your seed

"Write better code" ? bad seed. "Refactor the UserAuth class to use dependency injection and add unit tests" ? great seed. Specificity drives focus.

?

Say "reflect" at any time

Don't wait until the conversation ends. Mid-session reflections help you course-correct early and capture ideas before they're buried in long threads.

??

Save the seeds that emerge

The AI will label new ideas [SEED: ...]. Copy those into the Seed Reflection tool or your task manager. Those are your next projects.

Want SRT-1 always on?

Set up custom instructions in your AI settings so the protocol activates automatically — no thread starter needed. Works in Claude, ChatGPT, Gemini, and Copilot.