The public, one-way copy of the Claude Code harness shimo4228 uses every day: skills, subagents, rules and commit hooks you can lift one at a time, with the dated decision records behind each piece.
Agent Claude Code
Strict Architecture Decision Record reviewer. Reviews ADRs for section completeness, context-decision alignment, straw-man alternatives, one-sided consequences, unsourced numeric claims, and unstated override relationships with prior ADRs. Use PROACTIVELY after writing or substantially revising an ADR, before commit.…
Agent Claude Code
Context-starved prompt forager for diversity injection. Receives ONE line (a purpose) and deliberately nothing else, then searches external creativity-technique catalogs and prompt collections for 3-5 prompts from angles the requesting session would not produce itself. Dispatched by the prompt-perturb skill — not for…
Agent Claude Code
Generate concise, focused prompts using a lightweight model. Use when creating or rewriting LLM prompt templates.
Agent Claude Code
Expert Swift / SwiftUI code reviewer specializing in Swift 6 strict concurrency, value semantics, SwiftUI state management, memory management (retain cycles), and HIG compliance. Use for all Swift code changes in iOS / macOS projects.
Agent Claude Code
A research worker that investigates one assigned angle using web sources, registries, primary documents, papers, or a repository, then writes notes or a research report to a specified file. It is invoked by a lead process rather than starting research on its own.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: