ai-mktpl

Install and manage 1Password CLI (op) and op-exec in Claude Code sessions. Provides session-start auto-install for web sessions and comprehensive workflow skills.

Set tmux/iTerm2 tab titles to agent roles in agent team sessions. Disables Claude Code's LLM-generated titles and uses SessionStart hook to set role-based names.

Skills for configuring Claude Code, behavior correction, memory tracking, autonomy rules, and time-context awareness. Combines prompt history tracking, git-backed memory sync, and the self-checking Ralph loop pattern.

Skills for deploying docker-compose stacks via Arcane GitOps — directory conventions, secrets management with 1Password, and GitHub Actions CI/CD.

Skills for managing Cloudflare developer platform components including Workers, R2, D1, AI Gateway, Pages, Tunnels, and more — with Pulumi IaC setup for each.

Simplify and refine code for clarity, consistency, and maintainability. Requires pr-review-toolkit plugin.

Help claude understand, discover, and execute slash commands when accidentally sent as a message.

Common-sense rules for AI assistant behavior. On session start, symlinks bundled rules into the project's .claude/rules/common-sense directory so they are automatically loaded as context.

Prevents context window bloat from large tool outputs. PostToolUse hook detects oversized results and instructs Claude to save to file instead of keeping in context. PreToolUse hook redirects Bash commands that are likely to produce large output.

Command to correct AI behavior mistakes and update rules to prevent recurrence

Create and maintain Claude Code slash commands with guided assistance

Skill for generating daily organization-wide reports covering commits, PRs, branches, and issue changes across all repositories in a GitHub org. Designed for scheduled CI workflows and on-demand reporting.

Data format conversion and querying utilities. Supports YAML, JSON, TOON (Token-Oriented Object Notation), and XML/HTML transformations with special handling for Playwright accessibility snapshots. TOON provides 30-60% token reduction vs JSON for LLM contexts.

Multi-agent deep research system for complex investigations. Uses a lead-researcher orchestrator, parallel sub-researchers, and a critical-reviewer to produce evidence-based reports with confidence levels and citations.

File editing utilities: auto-formatting and linting for written/edited files with project-aware configuration

Relentlessly fix a PR until CI passes — iterates through review, fix, push cycles until green

Skill for managing stacked Git branches with git-spice (gs) CLI tool

GitHub CLI installation, authentication, and workflow skill for Claude Code sessions. Consolidates gh-tool and github-auth-skill into a single plugin.

Automatic GitHub App token lifecycle for Claude Code sessions. Generates installation tokens on session start, monitors expiry via PreToolUse hook, and refreshes transparently before commands that need authentication.

Install and manage the Google Workspace CLI (gws) with per-service skills for Gmail, Calendar, Drive, Docs, Sheets, Slides, Chat, Tasks, Contacts, and Admin

Skills for MCP CLI tools, frameworks, proxy daemoning, and gateway patterns. Covers philschmid/mcp-cli, mcp-use, apify/mcpc, mcp-proxy daemoning, and MCP gateway architecture for sharing servers across sessions and networks.

Intelligent memory management for CLAUDE.md files. Automatically detects and stores user preferences, rules, and instructions with smart scope detection and hierarchical organization.

Install and manage mise (tool version manager) in Claude Code sessions. Provides session-start auto-install for web sessions and comprehensive workflow skills.

Generate OpenGraph images for Nuxt/Vue pages using HTML templates and html2png.dev API

Merge permission scopes from configurable source settings.json files into settings.local.json on session start. Supports project-level and user-level installation.

Skill for installing, updating, and managing Claude Code plugins — including what requires a restart and what hot-reloads.

Skills for managing Proxmox VE hosts and LXC containers — creating containers, running Docker in LXC, and hosting services like cloudflared.

Sync upstream Claude config repo on session start. Pulls latest from a configured git repo to ~/.claude-remote/ and reports update status.

Review code changes with detailed feedback on quality, security, performance, and maintainability

Pre-tool-call safety evaluation using prompt-style hooks for inline AI evaluation

Pre-tool-call safety evaluation using Claude CLI (haiku) with script-style hooks

Source control management utilities for improving interactions with branches and PRs, both locally and in CI environments

Allows Claude to gracefully terminate itself using SIGINT

Set up the sequential-thinking MCP server and auto-configure permissions for its tools on session start.

Skill for generating structured session reports — fact-finding across git history, GitHub issues and PRs, tasks, and agent contributions to produce a comprehensive work summary. Works for solo agents and agent teams alike.

Enforces that specified skills are loaded before certain tools can be used. Tracks skill reads via PostToolUse and blocks tool use via PreToolUse if the required skill hasn't been recently loaded.

Agent skill for maintaining, updating, and improving existing Claude Code skills

Configurable status line for Claude Code showing session info, project context, and git status

Status line for Claude Code with iTerm2 badge integration - shows session info, project context, git status, and updates iTerm2 badge

Intelligent skill that helps Claude parallelize Task tool calls when working on repetitive or batch operations, optimizing throughput based on task complexity

Launch independent Claude sub-agents in tmux sessions with custom configurations, tool restrictions, and iTerm integration

Enforces commit-on-complete for tasks and reminds about ephemeral session awareness

Automatically syncs todos and plans from ~/.claude/ to the current project's .claude/ directory

Auto-approve Edit, Write, and Bash permission requests in Claude Code web sessions to suppress interactive permission prompts

Hook-based workflow for capturing unstructured thoughts and triggering the exec-assist agent to categorize them into actionable items (issues, tasks, docs).

Skills for using z.ai (Zhipu AI) GLM models and configuring Claude Code to access them via the Anthropic-compatible endpoint or API gateways.