Plugin Claude Code
Plugin marketplace listing 2 plugins: claude-hud, motif.
Plugin Claude Code
Plugin marketplace listing 2 plugins: claude-hud, motif.
Instructions file
Claude Code instructions for zackbart/agent-plugins, covering agent-plugins, structure, working on a plugin, adding a new plugin and version management.
Plugin Claude Code
Real-time statusline HUD for Claude Code - context health, tool activity, agent tracking, and todo progress.
Command
Configure HUD display options (layout, presets, display elements) while preserving advanced manual overrides.
Command
Configure claude-hud as your statusline.
Plugin Claude Code
Personal development workflow plugin. Encodes a 4-stage process (Research, Plan, Build, Validate) for any codebase.
Settings file Claude Code
Personal Claude Code plugin marketplace - claude-hud, motif, helm.
Agent
Write-capable build agent. Spawned during the motif workflow to execute specific implementation tasks in parallel. Receives task assignments and a shared context artifact. Reports back with results and issues.
Agent
Adversarial critic subagent. Receives a plan or proposal, actively inspects the codebase to verify claims, and returns a severity-ranked list of concerns. Read-only access. Spawned during the motif dev workflow after planning.
Agent
Deep codebase exploration agent. Spawned during the motif workflow to gather structured research before planning. Read-only — never modifies files. Self-calibrates depth based on task complexity.
Agent
Independent validation subagent. Spawned after the Build stage to audit completed work against the original task, plan, and codebase standards. Read-only access. Separate eyes from the builder.
Agent
Deep web research agent. Use when a task needs external knowledge — unfamiliar APIs, library migrations, architectural patterns, error diagnosis, technology comparisons, or any question where the codebase alone isn't enough. Goes deep: iterates across multiple search rounds, follows leads, cross-references sources…
Hook
Runs when a session starts, running an inline shell check (2 commands). From zackbart/agent-plugins.
Skill Claude CodeCodex
Consult the Codex CLI for a second opinion on any question — a sanity check, a code review, a design decision, a "what would another model say." Use when the user asks to "ask codex", "what does codex think", "get a second opinion", "cross-check with codex", "cross-check with gpt", or invokes /motif:ask-codex. Also…
Skill Claude CodeCodex
Personal development workflow orchestrator. Use when starting any development task — bug fixes, features, refactors, or explorations. Runs a 4-stage process: Research, Plan, Build, Validate. The plan is the single approval gate — once approved, everything else runs autonomously.
Skill Claude CodeCodex
Model-routing policy for orchestrating work across models (gpt-5.5, sonnet-5, opus-4.8, fable-5). Use when choosing which model to run a task, subagent, or workflow on — "which model for this", "route this work", "what should I delegate", "cheapest model that meets the bar", "which model is best at X" — or when the…
Instructions file CodexOpenCode
AGENTS.md instructions for zackbart/connecta, covering working on connecta, verification, the map, deployment setup and where new code goes.
Instructions file
Claude Code instructions for zackbart/connecta, a project described as: A single MCP endpoint that aggregates downstream connectors behind a compact meta-tool interface.
MCP server Claude CodeCodexCursor
One MCP to rule them all — a single MCP endpoint aggregating many downstream connectors behind a code-first surface of seven meta-tools. Runs locally from the @zackbart/connecta npm package.