app-vitals-marketplace

Structured orchestration for Ralph loops with PRD support, progress tracking, TDD guidance, and escape hatches. Supports implementation, documentation, and investigation task types. Wraps the official ralph-loop plugin.

Defense-in-depth protection for Claude Code. Plugin adaptation of IndyDevDan's claude-code-damage-control (https://github.com/disler/claude-code-damage-control). Blocks dangerous commands and protects sensitive files via PreToolUse hooks. Features pattern-based blocking, ask confirmations, and three-tier path protection (zero-access, read-only, no-delete).

Captures corrections, discoveries, and preferences as you work with Claude. Stores learnings in CLAUDE.md, promotes to global scope, and helps contribute improvements back to marketplace plugins.

Interactive PR review workflow with local drafts, CLAUDE.md compliance checking, and batch processing. Auto-detects first reviews vs. follow-ups via GitHub history. Supports iterating on reviews before posting to GitHub.

Browse and read meeting transcripts captured by the mt CLI tool. Provides a /transcripts command to list recent recordings and read specific transcripts by time reference.

Distill conversations into persistent context files. Extracts long-term facts, goals, and decisions into your project's context/ directory before clearing.

AI-powered triage of Dependabot PRs. Analyzes individual PRs and posts patrol-style risk assessments. Bulk command works through all open Dependabot PRs and gets them to merge-ready.

Shipwright — structured dev pipeline with plan sessions, task execution, autonomous dev loops, and multi-agent code review for any software project

Opinionated Terraform best practices for Claude Code. Automatically enforces exact version pinning, conservative tagging, pre-commit validation (fmt, validate, init, plan), Terratest generation, state isolation, module composition, variable validation, and naming conventions. Activates whenever working with .tf or .tfvars files.

Continuous code health enforcement via golden principles. Scans for drift (dead code, missing tests, TODO debt, inconsistent patterns, documentation gaps) and opens focused PRs to fix it. Designed to run as a daily cron or on-demand slash command.

Agent-readiness audit and bootstrapping for codebases. Scans for legibility gaps (missing CLAUDE.md, docs outside the repo, untested paths, no observability) and scores each category. Use --fix to generate the missing pieces so Claude can work effectively from day one.