pdr-rtv

Coordinate parallel build-phase patch rollouts with tournament-style winner selection.

1|Updated Mar 15, 2026
One-click install
npx skills add https://github.com/paulingham/.claude --skill pdr-rtv
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pdr-rtv
Source: https://github.com/paulingham/.claude/tree/main/skills/pdr-rtv
Command: npx skills add https://github.com/paulingham/.claude --skill pdr-rtv

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Parallel-Diverse-Refine + Recursive-Tournament-Verification (PDR-RTV) orchestrates scalable patch discovery during the build phase by coordinating parallel rollouts and tournament-style evaluation to surface high-quality patches.

Core Features & Use Cases

  • Parallel-dispatch across iterations to explore multiple patch candidates concurrently.
  • Deterministic refinement by sampling prior rollouts to guide subsequent evaluations.
  • Tournament-based patch-critic style decision making to select a winning patch and propagate it into the pipeline.

Quick Start

Invoke the pdr-rtv build-phase workflow to run two iterations with four parallel rollouts per iteration.

Frequently Asked Questions about pdr-rtv

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I coordinate parallel patch exploration during a build phase?

You can coordinate parallel patch exploration by running multiple concurrent rollouts across iterations, using deterministic sampling of prior rollouts to guide subsequent candidate evaluation and selection.

What is tournament-style selection for patch candidates?

Tournament-style selection is an evaluation mechanism where multiple concurrent rollout candidates compete per iteration, allowing a patch-critic approach to deterministically select and propagate a single winning patch into the pipeline.

How do I run iterative build-phase rollouts with deterministic sampling?

Iterative build-phase rollouts are executed using a two-iteration, four-rollout-per-iteration configuration, where deterministic sampling of prior rollout results guides the refinement of subsequent patch candidates.

Can I use this parallel dispatch workflow in an existing CI pipeline?

Yes, the parallel dispatch workflow is applied directly within CI/build pipelines, using frontmatter-defined dispatch rules to manage rollout summaries and tournament-driven winner selection across build iterations.

What are the limitations of a fixed two-iteration, four-rollout configuration?

The fixed two-iteration, four-rollout-per-iteration configuration limits exploration depth, meaning patch discovery is constrained to eight total rollouts, which may be insufficient for highly complex build-phase patch landscapes.