Plugins

from DataflightSolutions/claude-plugins

Persistent cross-device knowledge-graph memory (MCP + usage skill + setup command)

from titabash/claude-plugins

Graylog MCP server — search and monitor logs across eRegistrations instances with Elasticsearch syntax. Uses Graylog-native auth (NOT Keycloak). Requires bpa-mcp plugin for instance management.

GRC domain knowledge — 15 frameworks, 24 commands, cross-framework mapping, document review, and operational workflows. Cloud-agnostic.

Green Train dev-workflow skills: GitHub backlog governance as a manager loop (backlog-manager) — triage labels, append-only description completion, bounded ready queue (Todo ≤ 5) on a Projects board, board drift repair. Config-driven per repo via .claude/backlog-manager.yaml, DRY-RUN by default.; plus loop-or-not, a decision skill that inspects your repo and rules whether a task deserves an agent loop (don't loop / timer loop / goal loop) and drafts the contract

Green Train file skills: free 100GB+ disk space with safe cleanup, organize 1000+ files with Smart Folders, convert office documents to a searchable knowledge base, and orchestrate all three in one workflow

Green Train media skills: download videos from YouTube and 1000+ sites, download X/Twitter videos, download Apple Podcasts episodes, convert PDFs to images, generate viral video titles from subtitles, transcribe audio/video/URLs to txt/srt/vtt/json locally with mlx_whisper, local TTS/STT on Apple Silicon, generate visual decks on open-slide with Green Train's design system (visual-deck v1.0), and a deprecated batch-template Slides injector (visual-slides — narrow use only)

Green Train perspective skills: persona thinking-framework advisors distilled with nuwa-skill. Currently includes jordan-peterson-perspective (the 'lobster professor') — 6 mental models, 8 decision heuristics, full expression DNA, with built-in failure-mode annotations and safety boundaries.

Green Train planning skills: think-before-you-slide PPT methodology. Classify PPT type (pitch / research / teaching / narrative), set up research-driven thesis with six-question specificity diagnostic, and review storyline for structural fit. Pairs with visual-deck for the full pipeline.

Deep research superpowers for your AI agent. Three research tiers (quick ~25s, deep ~5min, ultra up to 1hr) powered by GREP.

AI-powered codebase search and understanding. Query your repositories using natural language to find relevant code, understand dependencies, and get contextual answers about your codebase architecture.

from jcaldwell-labs/my-grid

Dialectic claim discipline for AI agents: scholastic vocabulary, peer-grill file-based reconciliation, runnable verifiers/falsifiers, multi-agent ratification, fleet-ratify N-agent attestation, permutation NxN fleet topology ratification. Synthesis: an unexamined claim does not exist.

Interview the user relentlessly about a plan or design, resolving each branch of the decision tree until shared understanding. Use to stress-test a plan.

Interview the user relentlessly about a plan or design. Use when the user wants to stress-test a plan before building, or uses any 'grill' trigger phrases.

Relentless plan-and-design interrogator. Walks the decision tree one branch at a time, asking forcing questions sequentially with recommended answers. Explores codebase before asking. Derived from Matt Pocock's MIT-licensed grill-me with: (1) 3 stdlib Python tools (decision-tree extractor across 6 branch kinds, question generator with dependency-aware ordering, JSON-backed session tracker for multi-day grills), (2) 3 references citing 7-8 sources (6 forcing-question patterns, when to stop grilling, companion tooling), (3) cs-grill-master persona agent + /cs:grill-me slash command. Matt's relentless one-at-a-time interview discipline preserved verbatim per MIT.

Interview the user relentlessly about a plan or design until reaching shared understanding

Interview the user relentlessly about a plan or design, resolving each branch of the decision tree until shared understanding. Use to stress-test a plan.

Relentless plan-and-design interrogator. Walks the decision tree one branch at a time, asking forcing questions sequentially with recommended answers. Explores codebase before asking. Derived from Matt Pocock's MIT-licensed grill-me with: (1) 3 stdlib Python tools (decision-tree extractor across 6 branch kinds, question generator with dependency-aware ordering, JSON-backed session tracker for multi-day grills), (2) 3 references citing 7-8 sources (6 forcing-question patterns, when to stop grilling, companion tooling), (3) cs-grill-master persona agent + /cs:grill-me slash command. Matt's relentless one-at-a-time interview discipline preserved verbatim per MIT.

Stress-test your plan before vibe coding. The AI asks you questions to build a shared understanding — you answer in a sleek web UI.

Docs-anchored grilling session — interrogates a plan against the project's existing language (CONTEXT.md) and recorded decisions (docs/adr/), updating those files inline as terminology and decisions crystallise. Derived from Matt Pocock's MIT-licensed grill-with-docs with stdlib validators (CONTEXT.md linter, ADR scanner, glossary-code consistency), reference docs, cs-grill-with-docs agent, and /cs:grill-with-docs command.

Grilling session that challenges plans against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline

Plan-grilling and session-handover skills for project docs. grit-grill interviews you about a plan along the GRIT axes (Guts/Resilience/Initiative/Tenacity), ends with an ASCII radar scorecard, and crystallizes intent into docs/intent.md; handover compacts the session state into docs/handover.md.

A programming language for agentic finance - structured, reviewable spells that compile to deterministic execution with explicit AI judgment boundaries

Writing partner for novelists: surgical editing, cold-read audits, codex memory, voice-spec anchoring. Never writes your book for you.

Search local retailers, build hardware/shopping lists with real prices and stock

Use xAI's Grok from Claude Code: ask, research (effort=max + live web search + --check self-verification), review, adversarial-review, imagine, imagine-video, parallel best-of-N and fan-out, model picker, and long-form rescue via subagent. Wraps the local `grok` CLI (xAI Grok Build) with env scrubbing, ANSI sanitization, capability probing, TOCTOU/symlink defenses, prompt-injection defenses, and an optional stop-time review gate.

Use Grok Build from Claude Code to review code or delegate tasks.

Run grok reviews and delegate rescue work from Claude Code.