compound

Name: compound
Availability: InStock
Author: sr-ai-dev

Convert pull request discussions and spec context into dated Markdown learning documents.

Updated Apr 14, 2026

Search Tags:#documentation #pull request #github #knowledge extraction #review feedback #learning synthesis #problem classification

Ready to run

npx skills add https://github.com/sr-ai-dev/sr-harness --skill compound-sr-ai-dev

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: compound
Source: https://github.com/sr-ai-dev/sr-harness/tree/main/skills/compound
Command: npx skills add https://github.com/sr-ai-dev/sr-harness --skill compound-sr-ai-dev

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you capture and preserve valuable knowledge from completed pull requests so learnings don’t get lost in review threads.

Core Features & Use Cases

Extract PR context to locate the relevant spec plan and context directories for a consistent documentation target.
Aggregate knowledge sources by collecting context files and PR comments/reviews, then filter out noise to keep only actionable feedback.
Generate structured learning documentation using an internal template, classify the problem type, add tags, and create cross-references when related docs exist.
Use case: after merging a PR, automatically produce a dated, categorized learning record (with reviewer feedback) from the decisions/issues/learnings present in the spec context.

Quick Start

Run /compound with an optional PR number to generate a new document under docs/learnings/ from the PR and spec context.

Frequently Asked Questions about compound

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

FAQPage Schema

How do I extract learning documentation from GitHub pull request discussions?▼

To extract learning documentation from GitHub pull request discussions, run the /compound command with an optional PR number. It synthesizes reviewer feedback and PR context into a structured, dated Markdown file saved under docs/learnings/.

What is the best way to preserve reviewer feedback and PR context into reusable docs?▼

The best way to preserve reviewer feedback into reusable docs is to synthesize PR comments and spec context after completion. This process filters out noise, classifies the problem type, and generates a dated Markdown record with YAML frontmatter.

Does generating learning docs from pull requests require an existing spec plan?▼

Generating learning docs from pull requests relies on locating relevant spec plan and context directories. The Skill extracts these paths from the PR context to ensure the documentation target remains consistent and traceable.

How does knowledge extraction from pull requests handle duplicate documentation?▼

Knowledge extraction from pull requests handles duplicate documentation by deduplicating against existing docs in the target directory. It also adds cross-references when related learning records already exist to maintain a cohesive knowledge base.

When should I generate learning documentation from a pull request?▼

You should generate learning documentation from a pull request after PR completion or when you need to compound learnings from PR context. This ensures decisions, issues, and reviewer feedback are captured before they get lost in review threads.