# reflect

The reflect Claude Code prompt revises an optimizer package using the trace from a finished task. It turns observed evidence into rules for later tasks instead of carrying over one task’s settings.

- prompt · `type=prompt` · Claude Code, ChatGPT · ★ 99 repo stars
- Author: TokenRhythm
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/TokenRhythm/InfraRSI/tree/HEAD/infrarsi/rsi/a0/prompts/reflect.md
- Updated: 2026-11-12
- Security check: automated, 2026-11-17 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/prompts/reflect

## What it does

reflect edits an optimizer package after each development task. It receives the package and the trace from the task that just ended, with no other context. Its permitted changes cover skill guidance, search policy, role prompts, receipt-analysis scripts, and cross-task memory.

New guidance must connect conditions to observed evidence and account for counterexamples. Advice contradicted by the trace should be removed, and a configuration for one task must not become a general rule. The prompt restricts edits to the listed files and memory, then asks for a JSON response containing changed sections and a rationale.

## Install

### Claude Code (native platform)

**Manually: put the files in place**
- File: `.claude/commands/reflect.md`

### ChatGPT

**How to use**
- Paste the text at the start of a chat or into project instructions

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/TokenRhythm/InfraRSI/tree/HEAD/infrarsi/rsi/a0/prompts/reflect.md

## When to use it
- revise optimizer rules using evidence from a completed task
- remove advice that conflicts with the task trace
- record lessons and configurations for later tasks

## Example prompts
- Update the optimizer package using the latest task trace, and add only evidence-backed observations.
- Revise the skill rules to account for the trace’s counterexample and remove contradicted advice.
- Save the lesson from the completed task in cross-task memory and return the result as JSON.

## Similar
- [mantis-reflect](https://affpapa.org/en/nevibecode/skills/mantis-reflect.md) — skill · 1600 installs
- [neutral](https://affpapa.org/en/nevibecode/prompts/neutral.md) — prompt · ★ 480 repo stars
- [train-with-agent](https://affpapa.org/en/nevibecode/skills/train-with-agent.md) — skill · ★ 480 repo stars
- [dspy-optimizer-selection](https://affpapa.org/en/nevibecode/skills/dspy-optimizer-selection.md) — skill · ★ 125 repo stars
- [dspy-optimize-anything](https://affpapa.org/en/nevibecode/skills/dspy-optimize-anything.md) — skill · ★ 125 repo stars
- [nemo-mbridge-perf-cpu-offloading](https://affpapa.org/en/nevibecode/skills/nemo-mbridge-perf-cpu-offloading.md) — skill · 1800 installs

From the same set: [tokenrhythm/infrarsi](https://affpapa.org/en/nevibecode/by/tokenrhythm/infrarsi.md) — 19 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
