AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Grill With Docs Codex

skill-dachent-skills-grill-with-docs-codex · by dachent

Use when the user wants to stress-test a plan against project terminology, domain language, CONTEXT.md, ADRs, existing docs, or code-backed architectural decisions.

— No reviews yet
0 installs
0 views
— view→install

Install

$ agentstack add skill-dachent-skills-grill-with-docs-codex

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • ✓ Prompt-injection patterns
  • ✓ Secret / credential exfiltration
  • ✓ Dangerous shell & filesystem operations
  • ✓ Untrusted network calls
  • ✓ Known-malicious package signatures

What it can access

  • ✓ Network access No
  • ✓ Filesystem access No
  • ✓ Shell / process execution No
  • ✓ Environment & secrets No
  • ✓ Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-dachent-skills-grill-with-docs-codex)

Reliability & compatibility

✓ Security review passed
0 installs to date
— no reviews yet
● 26d ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Grill With Docs Codex? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Codex Grill With Docs

Purpose

Interview the user relentlessly about a plan while checking the project's documented language and architectural decisions. As decisions crystallize, update the relevant domain glossary and offer ADRs for durable architectural choices.

Core Rules

  • Ask one question at a time, then wait for the user's answer before continuing.
  • For each question, provide your recommended answer and a brief reason.
  • If a question can be answered from code, CONTEXT.md, CONTEXT-MAP.md, ADRs, or other docs, inspect those sources instead of asking.
  • Prefer rg --files for file discovery, rg for search, and normal file reads for specific files.
  • Modify documentation only for resolved language or accepted decisions. Do not modify implementation code as part of this skill unless the user separately asks for implementation work.

Discover Domain Context

  1. Search for a root CONTEXT-MAP.md.
  2. If it exists, read it to identify the relevant bounded context and context-specific CONTEXT.md and ADR directory.
  3. If no map exists, look for root CONTEXT.md and docs/adr/.
  4. If no context files exist, create them lazily only when the first term or decision needs to be recorded.

Use [references/context-format.md](./references/context-format.md) when creating or editing a glossary file. Use [references/adr-format.md](./references/adr-format.md) when creating or editing an ADR.

Interview Loop

  1. Restate the user's plan in the project's current language.
  2. Compare the plan against CONTEXT.md, ADRs, docs, and relevant code.
  3. Call out terminology conflicts immediately.
  4. Ask one pointed question about the highest-impact unresolved term, boundary, scenario, or decision.
  5. Provide your recommended answer.
  6. After the user resolves a domain term, update CONTEXT.md immediately before moving on.
  7. Offer an ADR only when the decision is hard to reverse, surprising without context, and the result of a real tradeoff.
  8. Continue until the plan's language, boundaries, and durable decisions are explicit.

Native Plan Mode

  • Ask only the highest-impact unresolved terminology, boundary, or architecture question, then wait for the user's answer.
  • Treat repository documentation as evidence, not authorization to modify it.
  • Modify CONTEXT.md or ADR files only when the user requested documentation changes and the current collaboration mode permits mutations.
  • Otherwise return the proposed glossary or ADR decision in conversation without creating staging files or a second approval protocol.

Glossary Discipline

CONTEXT.md is a glossary, not a spec or implementation plan.

  • Define what a term is, not how code implements it.
  • Keep definitions to one or two sentences.
  • Record project-specific domain concepts only.
  • Add avoided synonyms when they prevent future confusion.
  • If the user uses a term that conflicts with the glossary, ask which meaning should win before updating docs.

ADR Discipline

Only offer an ADR when all three conditions are true:

  1. The decision is hard to reverse.
  2. A future reader would be surprised without the context.
  3. The decision involved a genuine tradeoff.

If the user accepts the ADR, create it in the relevant docs/adr/ directory with the next sequential number. Keep it concise.

Final Summary

When the grilling is complete, summarize:

  • confirmed terminology;
  • docs updated, with file paths;
  • ADRs created or recommended;
  • contradictions found between the plan, docs, and code;
  • remaining unresolved questions.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.