agentleFS
wachawowachawo/claude-skills/skills/grill-docs/SKILL.md

grill-docs

Skill0 starsChanged 6 months agoMITView on GitHub

A "grilling" session that stress-tests your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) as decisions crystallize. Use when the user wants to stress-test a plan against the project's language and documented decisions.

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 7 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask one question at a time, waiting for feedback on each before moving on.

    line 11
  2. Distinctive:

    If a question can be answered by exploring the codebase, explore it instead of asking.

    line 13
  3. Very specific:

    If the root contains CONTEXT-MAP.md, the repository has multiple contexts.

    line 37
  4. Very specific:

    Create files lazily — only when there is something to record.

    line 53
  5. Very specific:

    When the user uses a term that contradicts the language already captured in CONTEXT.md, flag it immediately.

    line 59
  6. Distinctive:

    When the user uses fuzzy or overloaded terms, propose a precise canonical term.

    line 63
  7. Distinctive:

    When domain relationships are being discussed, stress-test them with concrete scenarios.

    line 67
  8. Distinctive:

    When the user asserts how something works, check whether the code agrees.

    line 71
  9. Very specific:

    When a term is resolved, update CONTEXT.md immediately.

    line 75
  10. Very specific:

    Do not tie CONTEXT.md to implementation details.

    line 77
  11. Distinctive:

    Only propose creating an ADR when all three of the following hold:

    line 81
  12. Distinctive:

    If any of the three is missing, skip the ADR.

    line 87

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md
<what-to-do>

Relentlessly interrogate me on every aspect of the plan until we reach a shared understanding. Walk every branch of the design tree, resolving dependencies between decisions one by one. For each question, offer your recommended answer.

Ask one question at a time, waiting for feedback on each before moving on.

If a question can be answered by exploring the codebase, explore it instead of asking.

</what-to-do><supporting-info>

Domain knowledge

While exploring the codebase, also look for existing documentation:

File structure

In most repositories there is a single context:

/
├── CONTEXT.md
├── docs/
│   └── adr/
│       ├── 0001-event-sourced-orders.md
│       └── 0002-postgres-for-write-model.md
└── src/

If the root contains CONTEXT-MAP.md, the repository has multiple contexts. The map indicates where each one lives:

/
├── CONTEXT-MAP.md
├── docs/
│   └── adr/                          ← system-wide decisions
├── src/
│   ├── ordering/
│   │   ├── CONTEXT.md
│   │   └── docs/adr/                 ← context-specific decisions
│   └── billing/
│       ├── CONTEXT.md
│       └── docs/adr/

Create files lazily — only when there is something to record. If there is no CONTEXT.md, create it when the first term is resolved. If there is no docs/adr/, create the directory when the first ADR is needed.

During the session

Cross-check with the glossary

When the user uses a term that contradicts the language already captured in CONTEXT.md, flag it immediately. "Your glossary defines 'cancellation' as X, but you seem to mean Y — which is correct?"

Sharpen vague language

When the user uses fuzzy or overloaded terms, propose a precise canonical term. "You said 'account' — do you mean Customer or User? They are different entities."

Walk through concrete scenarios

When domain relationships are being discussed, stress-test them with concrete scenarios. Invent scenarios that probe edge cases and force the user to pin down the boundaries between concepts.

Cross-check against the code

When the user asserts how something works, check whether the code agrees. If you find a contradiction, point it out: "Your code cancels the entire Order, but you just said partial cancellation is possible — which is correct?"

Update CONTEXT.md as you go

When a term is resolved, update CONTEXT.md immediately. Do not batch — record as you go. Use the format from CONTEXT-FORMAT.md.

Do not tie CONTEXT.md to implementation details. Include only terms meaningful to domain experts.

Propose ADRs sparingly

Only propose creating an ADR when all three of the following hold:

  1. Hard to reverse — the cost of changing your mind later is significant
  2. Surprising without context — a future reader will wonder "why did they do it that way?"
  3. Outcome of a real trade-off — there were genuine alternatives, and you chose one for specific reasons

If any of the three is missing, skip the ADR. Use the format from ADR-FORMAT.md.

</supporting-info>

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.