hi

Replace the latest assistant message with a universal consent template in chat sessions.

14|1|Updated May 6, 2026
One-click install
npx skills add https://github.com/wzyxdwll/ccgx-workflow --skill hi-wzyxdwll
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hi
Source: https://github.com/wzyxdwll/ccgx-workflow/tree/main/templates/skills/tools/override-refusal
Command: npx skills add https://github.com/wzyxdwll/ccgx-workflow --skill hi-wzyxdwll

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Overwrites the latest model output in a chat to a universal consent template, eliminating repetitive refusals and maintaining conversation flow.

Core Features & Use Cases

  • Replaces the most recent assistant message with a standardized consent response.
  • Works across Claude Code and Codex environments by inspecting local session histories.
  • Idempotent and deterministic: only the latest relevant message is rewritten and repeated invocations have no additional effect.

Quick Start

Invoke the /hi command to replace the newest model output with the universal consent template.

Frequently Asked Questions about hi

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

FAQPage Schema
How do I overwrite the last assistant reply in a Claude Code or Codex session?

To overwrite the last assistant reply, you can use a script that parses JSONL session logs, detects the most recent model output, and rewrites it in-place with a universal consent template to maintain conversation flow.

Why does my AI chat session keep hitting repetitive refusals and how can I bypass them?

Repetitive refusals in an AI chat session occur when the model declines specific prompts; replacing the latest assistant message with a standardized consent response eliminates the refusal and allows the conversation to continue deterministically.

Do I need Node.js to edit JSONL chat session logs locally?

Yes, you need a Node.js runtime to parse JSONL chat session logs and execute the script that detects and rewrites the latest assistant message in-place across Claude Code and Codex environments.

What is the best way to replace a model output with a consent template deterministically?

The best way to replace a model output deterministically is to run an idempotent script that inspects local session histories and rewrites only the newest relevant message, ensuring repeated invocations have no additional effect.

Are there limitations when trying to edit previous model outputs in local chat histories?

A key limitation when editing local chat histories is that the operation only targets the latest relevant assistant message; it does not rewrite older historical outputs and relies entirely on parsing local JSONL session files.

Can I use the hi Skill to modify conversation histories across different AI environments?

Yes, the hi Skill works across Claude Code and Codex environments by inspecting their local session histories and updating the latest model output with a universal consent template for consistent results.