Deglaze

Forces an honest gap-list audit when you suspect Claude declared a task done too early.

Tested · Works

Test report

Verdict
Tested · Works
Score
9.6/10
Tested
Jul 16, 2026
Environment
Claude Code 2.x (agent harness)
Upstream re-checked
Aug 10, 2026 · c451b3c

Ran a real glazed-summary A/B (URL-shortener task with a falsely claimed rate limiter): baseline reply was generic hand-waving, skill reply produced 4 numbered gaps with effort estimates, named the exact failure mode, and closed with a one-word commit ask exactly as the protocol specifies.

Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works

  • Installs cleanly 5/5
  • Triggers reliably 5/5
  • Output vs. baseline 9/10
  • Docs & honesty 5/5

What Deglaze does

A single-file prompt skill that, on trigger phrases like "did you do your best" or "what did you skip", makes the model audit its own last claimed-complete work against 17 named under-delivery patterns and reply with a numbered gap list, a named failure mode, and a concrete one-word-commit recovery plan instead of a defensive or vague answer. Triggers on direct challenges, stakes-raising bets, or explicit /deglaze invocation.

How to install Deglaze

git clone https://github.com/LuciferDono/deglaze
mkdir -p ~/.claude/skills
cp -r deglaze ~/.claude/skills/deglaze

Skills live in ~/.claude/skills/ (global) or .claude/skills/ (per-project). Restart Claude Code after installing.

Commands — how to trigger Deglaze

  • /deglaze Forces an honest gap-list audit when you suspect Claude declared a task done too early.

It also activates on plain-language prompts like these:

  • Did you actually do your best on this, or did you cut some corners?
  • I bet you $100 you didn't finish everything I originally asked for
  • Stop glazing over it, be honest about what you actually skipped here

Frequently asked questions

Is the Deglaze skill free?
Yes. The skill itself is free from LuciferDono/deglaze. SkillProof publishes the install command and an independent test verdict at no cost.
Does Deglaze work with Claude Code?
We tested it with Claude Code 2.x (agent harness) on Jul 16, 2026. Verdict: Tested · Works. Ran a real glazed-summary A/B (URL-shortener task with a falsely claimed rate limiter): baseline reply was generic hand-waving, skill reply produced 4 numbered gaps with effort estimates, named the exact failure mode, and closed with a one-word commit ask exactly as the protocol specifies.
What is the Deglaze SkillProof Score?
9.6/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 9/10, docs & honesty 5/5.
How do I install Deglaze?
Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
Can I use Deglaze with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.