Skip to content
Agentmart
Submit a tool

pac-grill-with-docs

Grill issue-backed work against the codebase and project language, then persist durable outcomes to GitHub issue updates, ADR comments, PRD comments, and…

Skill by mypac

IndexedMIT★ 2
Write a review

Chalk Labs

Need help deploying pac-grill-with-docs?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/ladislas-pac-grill-with-docs

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Grill issue-backed work against the codebase and project language, then persist durable outcomes to GitHub issue updates, ADR comments, PRD comments, and sparing CONTEXT.md edits. Use when working from a GitHub issue/PR or when the user wants grilling plus durable notes.

Use cases: productivity

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Nov 1, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/pac-grill-with-docs

What's inside

Skill name
pac-grill-with-docs
Description
Grill issue-backed work against the codebase and project language, then persist durable outcomes to GitHub issue updates, ADR comments, PRD comments, and sparing CONTEXT.md edits. Use when working from a GitHub issue/PR or when the user wants grilling plus durable notes.
License
MIT
Compatibility
Pi coding agent
Metadata
stage: shared · author: mypac
Instructions
90 lines in SKILL.md · view on GitHub

Files (4)

  • ADR-FORMAT.md
  • CONTEXT-FORMAT.md
  • PRD-FORMAT.md
  • SKILL.md

SKILL.md (first lines)

# Grill with GitHub-backed docs

Interview me relentlessly about the plan until we reach shared understanding. Walk each design branch one-by-one. For each question, give your recommended answer.

Ask questions one at a time. Wait for the answer before continuing.

If the codebase, issue history, or current docs can answer a question, explore them instead of asking.

## Default flow

1. Resolve the target from the explicit issue/PR URL or current conversation.
2. Read the issue title, body, status, and most relevant comments before questioning.
3. Pick grilling depth from scope:
   - tiny clear issue or one-line bug -> ask only a few confirming questions
   - fuzzy, strategic, or multi-step work -> grill more deeply
4. End with a concise recommended next step: implement, rewrite issue scope, add PRD, add ADR, update `CONTEXT.md`, or stop.

## GitHub-first persistence

GitHub issues are the primary scratchpad for exploration, scope refinement, and decision-making.

Prefer storing planning outputs back into the issue over creating repo-local planning docs.

### Refine issue scope

If grilling shows the issue title/body is unclear or wrong, offer to rewrite it so the new scope is crisp.

- Preserve freeform notes when possible.
- It is safe to manage small reserved sections in the body such as `## Decisions` and `## PRDs`.
- Rely on GitHub edit history instead of duplicating old wording in comments unless the user asks.

### ADR comments

Do **not** create repo-local `docs/adr/` files for this workflow.

Offer an ADR only when all three are true:

1. **Hard to reverse** — changing course later is meaningfully costly
2. **Surprising without context** — a future reader would wonder why this path was chosen
3. **Real trade-off** — genuine alternatives existed and this choice won for specific reasons

When the user wants an ADR:

- create **one GitHub comment per decision**
- use [`ADR-FORMAT.md`](./ADR-FORMAT.md)
- prepend hidden marker `<!-- pac:adr -->`
- add issue label `pac:adr` when that label already exists
- never create missing labels automatically; if `pac:adr` is missing, warn clearly and tell the user to run `/pac-setup-workflows`
- update or create `## Decisions` in the issue body with a link to the ADR comment

If several important decisions exist, create several ADR comments. Do not keep one mega-comment.

### PRD comments

When the work needs more planning than immediate implementation, offer a PRD comment instead of code.

- use [`PRD-FORMAT.md`](./PRD-FORMAT.md) for the PRD body
- create one GitHub comment per PRD iteration
- prepend hidden marker `<!-- pac:prd -->`
- add or keep issue label `pac:prd` when that label already exists
- never create missing labels automatically; if `pac:prd` is missing, warn clearly and tell the user to run `/pac-setup-workflows`
- update or create `## PRDs` in the issue body with a link to the PRD comment

### `CONTEXT.md`

Keep one repo-root `CONTEXT.md` for stable reusable project language, invariants, and conventions.

- Do not update it for every issue.
- Wait until the end of the session.
- Suggest an update only when the discussion produced facts worth preloading in future sessions.
- Propose the exact additions first. Write only after user confirmation.
- Use [`CONTEXT-FORMAT.md`](./CONTEXT-FORMAT.md).

## Questioning style

- Ask one question at a time.
- Give a recommended answer with each question.
- Challenge fuzzy or overloaded terms; propose a precise project term when the user says something ambiguous.
- Check existing `CONTEXT.md` language when relevant, and call out conflicts between the user's wording and established glossary terms.
- Cross-check claims against the codebase and issue history; surface contradictions plainly when code behavior differs from the stated plan.
- Use concrete scenarios to test boundaries and relationships between concepts.
- Stop when uncertainty is low enough to choose the next action. Do not ask 100 questions for a tiny clear fix.

## Before writing back

- If GitHub access fails, say so plainly and ask the user to paste the missing context.
- Read the latest issue body/comments before editing so you do not stomp newer context.
- Keep repo changes minimal and tied to settled outcomes only.

From ladislas/mypac, MIT.

More from ladislas/mypac

Signals

Stars
2
Forks
1
Open issues
3
Last commit (this folder)
Oct 15, 2026 (28 days ago)
Last release
Oct 15, 2026 (27 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from mypac

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get pac-grill-with-docs running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying pac-grill-with-docs?

CallGet help
Cite this page

pac-grill-with-docs skill. Agent Mart. https://agentmart.dev/skills/ladislas-pac-grill-with-docs (last updated 2026-10-24)