KairaxisKairaxis
Freestructured workflowlow risk

Capability · rank 07

grill-me

Get sharp, concrete critique that improves implementation quality.

Reviews implementation work with sharp, concrete critique and improvement suggestions.

Listed owner

mattpocock/skills

Package source

GitHub

Category

Developer Exp

Package validatedCreator reviewedDeveloper Exp
$npx skills add mattpocock/skills --skill grill-me
OverviewSandboxUsageExamplesLimitationsVersionsRelated Skills

Overview

Reviews implementation work with sharp, concrete critique and improvement suggestions. Review the sandbox comparison, limitations, permissions, and pinned version before install.

Outcome

Get sharp, concrete critique that improves implementation quality.

Who it is for

Engineers using agents for implementation review.

What it is not for

Soft encouragement or style-only nitpicks without substance.

Runtime export

codex_skill · markdown_export

Sandbox comparison
Generic baseline vs grill-me on the same sample input · pinned v1.2.0

Sample input

PR adds a sandbox runner with no timeout, retries, or output size limits.

Model: gpt-4.1-mini · temp 0.2

Generic baseline

38/100
Looks fine overall. Maybe add some error handling.

Skill-enhanced

93/100
Hard questions:
1. What happens on hung model calls?
2. How do you bound token/output cost?
3. Which failures are excluded from benchmarks?
Required fixes: timeout, output cap, retry policy, failed-run exclusion from lift metrics.

Rubric scores

+55 lift
Coverage of failure modes2 → 5
Actionable recommendations3 → 5
Measurable next steps2 → 4
Risk / limitation awareness2 → 4

Assumptions

  • Sample input is representative of a real production task
  • Same model and temperature for baseline and skill-enhanced runs

Comparison runs through the shared sandbox engine on registry sample cases (deterministic v0 path).

Usage

  1. 1. Inspect trust badges, limitations, and permissions on this page.
  2. 2. Run the sandbox comparison to see baseline vs skill-enhanced output.
  3. 3. Install the pinned version with the command below for your agent runtime.
  4. 4. Agents can also fetch /api/v1/skills/grill-me/manifest.
$npx skills add mattpocock/skills --skill grill-me

Example prompts & use cases

1 / Prompt

Grill this PR for missing edge cases

2 / Prompt

Challenge this architecture decision record

3 / Prompt

Stress-test this API contract

Limitations

  • Critique quality depends on provided context and constraints.
  • Does not run tests unless the runtime separately executes them.

Audience

Engineers using agents for implementation review.

Not for

Soft encouragement or style-only nitpicks without substance.

Version history

v1.2.0

Jan 16, 2026current

Stronger failure-mode grilling