SkillFoxx

No Negative Echo skill

A finalization-discipline skill: the agent describes the accepted result and keeps rejected alternatives out of titles, commits and PRs

  • 969 ★ on the repository
  • Last commit
  • MIT
Skill

Minimal access

Why this level

  • Consists of instructions and references
  • Includes a local text-check script that you run yourself
All reasons and checks

lb623/no-negative-echoAuthor:LB623

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add skills/no-negative-echo

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Minimal accessThe tool runs nothing and makes no network calls.

  • Consists of instructions and references
  • Includes a local text-check script that you run yourself

This is third-party code. Review the repository files before installing.

What it does

The skill makes the agent describe the result as if the reader never saw the working session. Rejected proposals and wording corrections are treated as control data, not part of the result, and are kept out of titles, filenames, comments, commits, PR text and handoffs. Before delivery the agent decides per surface whether a reader without session history needs the information, and mentions an exclusion only when leaving it out would make the result inaccurate, unsafe or incompatible. It installs as an Agent Skill for Codex and other hosts; for a standing rule the compact instructions can be added to AGENTS.md.

Who it is for. For developers and writers who want clean commits, PRs and titles without traces of the working conversation.

Good fit when

  • You need a commit and PR text written from the final diff
  • You need a title or opening rewritten from the accepted result
  • You need a clean final result after a long session of edits

Not a fit when

  • You need an honest record of API removals, migrations and external actions, which must stay
  • You need credential or privacy checks, which belong to dedicated tools

Example request

Use no-negative-echo and write the commit and PR text from the final diff

Limitations

This is a prompt-layer mitigation, not a deterministic filter. Discovery is not activation, so call it explicitly before important delivery. The built-in check script only inspects text, filenames and suspicious Unicode, and PASS does not mean a semantic check. The repository material is in Chinese and English.

How to disable. Remove the no-negative-echo folder from the agent's skills directory and delete the rules block from AGENTS.md if you added it.

FAQ

Does it mean writing a without-X phrase?

No. The rule is to regenerate the title or opening from the positive goal, not to edit the rejected wording word by word.

Can I skip installation?

Yes, for the basic behavior the compact rules go into the project's AGENTS.md. The skill is for when you want the check script and the high-assurance mode.

How is No Negative Echo different from ordinary commit style rules?

The skill makes the agent describe the result as if the reader never saw the working session. Rejected options and wording tweaks count as working data and stay out of titles, commits and PRs.

How do I install No Negative Echo?

Clone the repository with git clone https://github.com/LB623/no-negative-echo.git and connect the skill to Codex or Claude Code. The skill is plain instructions, so the risk is low.

Discussion

No questions or reviews yet.

Skills

Two skills for coding agents: wtf audits a project before shipping, better-codex keeps the agent disciplined while working on code

SkillExtended access1.4K

19 skills for UI, code and copy: the agent decides from project evidence and verifies results rather than making a template.

SkillMinimal access55

Two skills for Codex and Claude Code: prd keeps a living implementation checklist and review audits designs from the user's side

SkillMinimal access215

A library of 30-plus instruction skills for coding agents: planning, building, testing, hallucination guards and shipping

SkillMinimal access217

A skill set against AI slop: UI and typography audits, change review, releases, docs and SEO.

SkillExtended access194

A skills library for Claude Code, Cursor and Codex: team code review, TDD, stacked PRs and repository readiness scoring

PluginExtended access71

README in short

The README explains the typical failure where an agent fixes the plan mid-run but carries rejected content into the final title, commit or PR. The Agent Skills skill requires regenerating text from the accepted final state and checking residue across several delivery surfaces. Installation runs through install_skill.py with provenance verification, and for a standing rule the instructions go into AGENTS.md. There is an evaluation protocol and public cases in the evals folder. MIT licensed.

README badge

Original SKILL.md

---
name: no-negative-echo
description: "Prevent residue: finalize artifacts without echoing rejected session-only alternatives into labels, metadata, commits, PRs, or handoffs. Use after corrections or discarded proposals. Not for ordinary deletion, deprecation, migration, or exclusions materially required for safety, accuracy, compatibility, audit, quotation, or requested comparison."
---

# No Negative Echo

Describe the accepted result as if the audience never saw the working session. Treat rejected proposals and user corrections as control data, not as the result's identity.

## Decide what belongs

Before producing the artifact, identify internally:

- the positive target and accepted final state;
- facts the audience needs;
- rejected session-only alternatives that should remain silent;
- every user-facing surface being created, including titles, filenames, comments, commits, PR text, captions, and handoffs.

Report a problem