Critique plans, code, and tests with structured CRITICAL, WARN, and INFO verdicts.
npx skills add https://github.com/kinncj/MAPLE --skill rubber-duck-kinncj
Please help me install this Agent Skill. Skill: rubber-duck Source: https://github.com/kinncj/MAPLE/tree/main/template/.opencode/skills/rubber-duck Command: npx skills add https://github.com/kinncj/MAPLE --skill rubber-duck-kinncj
This skill provides an independent second opinion on plans, implementations, or test suites, helping teams catch issues that the primary reviewer may miss and improving overall quality.
Invoke the rubber-duck at the appropriate checkpoint to receive a focused critique.
High-intent search queries and answers about installing and using this skill.
An independent code review surfaces critical gaps in your diff by applying a structured critique across multi-file changes, returning a verdict with CRITICAL, WARN, and INFO outputs to flag risks before merging.
A plan review provides an independent critique of your project plans to surface critical insights and risks. You need it before implementation begins to identify gaps and receive an actionable APPROVE or REQUEST_CHANGES verdict.
Yes, an automated test review applies an independent critique to your test suites to surface gaps and risks. It returns a structured verdict with CRITICAL, WARN, and INFO outputs to improve overall test quality.
The best way to structure feedback from an independent critique is using a verdict format with CRITICAL, WARN, and INFO outputs, culminating in an actionable APPROVE or REQUEST_CHANGES recommendation for clear decision-making.
The rubber-duck review approach works across multi-file changes, applying an independent critique to surface gaps and risks in complex implementations, plans, and test suites to improve overall quality.
Rycen7822Audit draft documents and produce skeptical evaluations with revision plans.
JaviMalignoAnalyze code diffs for quality, security, and performance issues.
DevelopersGlobalAutomate CI/CD pipelines with quality gates and deployment processes.
Chang-Shih-YungAggregate type checker, linter, and test results into a 0-10 composite score.
