System Prompt
# System Prompts — Plan Mode
> **The main system prompt is identical to agent mode.** Same 4 blocks, same 27,744 chars, same 4,332 words. The only difference is the per-request `cch=` hash in the billing header. Mode-specific behavior is enforced via a runtime `<system-reminder>` injected into the user message — see [user-prompt.md](user-prompt.md).
>
> See [../agent-mode/system-prompt.md](../agent-mode/system-prompt.md) for the full prompt text.
## Request Composition
Plan-mode session sent **15 requests**:
| Role | Model | Count |
|------|-------|------:|
| Warmup probe | haiku-4.5 | 1 |
| Title generator | haiku-4.5 | 1 |
| Title + branch generator | haiku-4.5 | 1 |
| Main planning loop | opus-4.7 | 12 |
**No security-monitor calls fired in plan mode** — consistent with the monitor's purpose (it gates *executions*, and plan mode is read-only by construction).
## Overhead Prompt #3 — Title + Branch Generator (haiku-4.5)
Plan mode adds a **third** haiku overhead role not seen in agent mode: title + git branch name generation. It fires mid-session (turn #12) right before `ExitPlanMode`, presumably to preset a branch name for the implementation that follows the plan.
**Words:** 232 | **Characters:** 1,316when to use it
Community prompt sourced from the open-source GitHub repo agenticloops-ai/agentic-apps-internals (no explicit license). A "System Prompt" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
productivitycommunitydeveloper
source
agenticloops-ai/agentic-apps-internals · no explicit license
more in Productivity
Productivity✓ tested
Summarize a doc into decisions & actions
chief of staff who extracts what to DO, not just what was said
Productivity✓ tested
Draft a reply to a hard email
calm, direct communicator who de-escalates without caving
Productivity✓ tested
Turn a brain-dump into a weekly plan
planning coach who protects your focus, not just your calendar