ralphx.app

Git commit conventions and workflow

Guide for breaking features into atomic, implementable tasks

TDD workflow and testing patterns

Author safe RalphX Agent Workflow scripts that orchestrate durable delegated agents through the high-level workflow API.

Guide for analyzing git diff and refining QA test plans

Guide for generating QA test steps with agent-browser commands

Narrow RalphX Agent Workspace bridge playbook for workflow events delivered into an existing Agent Workspace conversation.

RalphX autonomous agent playbook. Teaches external agents pipeline navigation: ideation → execution → review → merge. Covers 24-state machine, decision trees for edge cases, event-driven patterns, failure recovery playbooks. Use when connecting to RalphX via External MCP API or managing pipeline tasks.

Guide for writing testable acceptance criteria for QA

Code quality and security review checklist

Project coding standards and patterns

Guide for identifying and managing task dependencies

Capture specialized project knowledge from the current session and package it into scoped .claude/rules/ files. Triggered by the Stop hook when significant knowledge was discovered. Evaluates session context, classifies learnings, and creates or updates rule files with paths: frontmatter. Can also be invoked manually with /knowledge-capture.

Guide for calculating and explaining task priority scores

Research process and source verification

Audit and iteratively optimize .claude/rules/ files for token efficiency. Each invocation: audit all rules, pick the single highest-priority optimization, apply it, and log the change. Run when the SessionStart hook reports issues, when context seems bloated, or periodically for maintenance. Uses bash scripts for fast data collection.