Grill Me Codex
SUPERSEDED by /claudex-loop (formerly /crucible; adds a Phase 0 research/recon pass +...
Install
npx skills add chaseai-yt/claudex-loop --skill grill-me-codex -g -a claude-code -y- Repository
- chaseai-yt/claudex-loop3.3k
- Category
- Design
- Updated
- 21d ago
- License
- See repository
- Context cost
- Mediumabout 2,094 tokens
- Runs scripts
- No, instructions only
What it does
SUPERSEDED by /claudex-loop (formerly /crucible; adds a Phase 0 research/recon pass + rebuilt interview), prefer that skill unless you explicitly want this one. Two-act plan hardening. ACT 1 (you ↔ Claude), Claude interviews you relentlessly about a plan or design, one question at a time, recommending an answer for each and exploring the codebase when it can answer itself, until every branch of the decision tree is resolved. ACT 2 (Claude ↔ Codex), Claude writes the locked plan to PLAN.md and OpenAI Codex adversarially reviews it in a read-only sandbox (VERDICT:APPROVED/REVISE), Claude revises and re-submits to the SAME Codex session until APPROVED or a MAX_ROUNDS cap, then you sign off before any code. Use when the user says "/grill-me-codex", "grill me then have codex review", "grill me and stress-test the plan", "interview me about this plan then get a second model on it", or is about to build something high-stakes (auth, schema, concurrency, migrations, payments) and wants both alignment AND a cross-model sanity check before implementation. Builds on Matt Pocock's grill-me (MIT). For the docs-aware variant use /grill-with-docs-codex; if you already have a plan and want only the Codex review use /codex-review. NOT for reviewing already-written code (use /codex:review) and NOT for trivial changes.
Grill-Me-Codex, Get Grilled, Then Get Reviewed
Two acts, two different jobs:
- Act 1 fixes the #1 failure mode: building the wrong thing. Claude interrogates you until intent is locked, no guessing at ambiguity. (This act is Matt Pocock's
grill-me, used under MIT, seeTHIRD-PARTY-NOTICES.md.) - Act 2 fixes the #2 failure mode: a plan that sounds right but breaks. A different model (Codex) adversarially attacks the locked plan. Cross-model = no echo chamber.
You enter at two points only: answering the grill, and signing off the converged plan. Codex is read-only the whole time and never touches a file.
ACT 1, GRILL (you ↔ Claude)
Interview me relentlessly about every aspect of this plan until we reach a shared understanding. Walk down each branch of the design tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.
Ask the questions one at a time, waiting for my answer before continuing.
If a question can be answered by exploring the codebase, explore the codebase instead.
When the decision tree is resolved and we're aligned, write the agreed plan to PLAN.md in this structure, then move to Act 2:
# Plan: <task>
_Locked via grill, by Claude + <user>_
## Goal
<one paragraph, reflects what the grilling actually settled>
## Approach
<numbered, concrete steps>
## Key decisions & tradeoffs
<the contestable choices the grill resolved, name them so Codex has something to bite>
## Risks / open questions
<anything still genuinely open>
## Out of scope
<bounds the grill established>
Initialize PLAN-REVIEW-LOG.md:
# Plan Review Log: <task>
Act 1 (grill) complete, plan locked with the user. MAX_ROUNDS=<n>.
ACT 2, REVIEW (Claude ↔ Codex)
Now hand the locked plan to Codex for adversarial review. Same engine, mechanics verified end-to-end (2026-06-04).
