Bounded Prompt History

Bound conversation history with a configurable cap and newest-first ordering.

Updated Mar 13, 2026
One-click install
npx skills add https://github.com/JDL440/nfl-eval --skill bounded-prompt-history
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: Bounded Prompt History
Source: https://github.com/JDL440/nfl-eval/tree/main/.squad/skills/bounded-prompt-history
Command: npx skills add https://github.com/JDL440/nfl-eval --skill bounded-prompt-history

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Unbounded runtime prompt history can cause prompts to grow without bound, increasing latency and making it harder to reason about context.

Core Features & Use Cases

  • Cap and trim editor history to a fixed size, preserving the most recent turns.
  • Apply deterministic ordering and safe formatting to keep prompts predictable across runs.
  • Use cases include editor prompts, reviewer notes, and any multi-turn context where recency matters but total history must remain bounded.

Quick Start

Configure a MAX_EDITOR_PREVIOUS_REVIEWS cap and enable newestFirst in getArticleConversation to bound prompt history.

Frequently Asked Questions about Bounded Prompt History

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I cap runtime prompt history to prevent unbounded conversation growth?

You can cap runtime prompt history by applying a constant limit to trim older turns, preserving the most recent conversation context while preventing unbounded prompt growth and increasing latency.

What is bounded prompt history and when do I need it for multi-turn context?

Bounded prompt history limits the total conversation turns retained in runtime context. You need it for multi-turn contexts like editor prompts and reviewer summaries where recency matters but total history must remain constrained.

How do I configure newest-first ordering for deterministic conversation queries?

Configure newest-first ordering by enabling the newestFirst parameter in your conversation query function, then apply deterministic sorting in the formatter to ensure prompts remain predictable across runs.

Does bounded prompt history work with editor prompts and reviewer summaries?

Yes, bounded prompt history works with editor prompts, reviewer summaries, and any multi-turn context where recency matters. It applies a cap constant and deterministic formatting to keep prompts predictable across runs.

What are the limitations of trimming conversation history to a fixed size?

Trimming conversation history to a fixed size discards older turns to prioritize recent context. This approach loses earlier conversation details, making it unsuitable when comprehensive historical reasoning across all turns is required.