Fable's engineering discipline for any Claude session, with one habit enforced by a hook. Two layers: running the task (stage, verify each stage with a failable check, written done-criteria) and writing the code (recon before edit, verify against a copy with a negative self-test, single-writer chokepoints…
★not rated 5 2mo agoA
tokens not measured
originalMIT
Verification-first discipline for coding agents: a 9-rule doctrine, 11 skills, a risk-routed adversarial reviewer, and fail-open enforcement hooks. Done is proven by a command, not by opinion.
★not rated 5 11d agoA
tokens not measured
originalMIT
A live, auto-refreshing per-session summary brief docked beside your Claude Code terminal — to quickly get up to speed when switching between parallel sessions.
★not rated 5 5d agoA
tokens not measured
originalBSD-3-Clause
Automatically compress large natural text or log files before processing using LLMLingua to save tokens. Trigger when the user pastes massive text blobs, asks to analyze a large file (logs, docs, transcripts), or provides a prompt too large for the context window.
★not rated 5 4mo agoC
tokens not measured
originalMIT
End-to-end pipeline: pull a task from Redmine by ID, implement it with the Claude CLI, verify the result in a real browser via the claude-in-chrome extension, and keep the Redmine ticket in sync throughout.
★not rated 5 10d agoA
tokens not measured
originalMIT
Skills for working with Hookdeck Event Gateway (receive, queue, route, and deliver webhooks) and Hookdeck Outpost (send webhooks and events to user-preferred destinations). Includes a router skill that delegates to event-gateway or outpost based on the task.
★not rated 5 16d agoA
tokens not measured
originalMIT
Auto-wrap any Click/typer CLI as an MCP server — zero-code transformation. Supports stdio, HTTP+SSE, and Streamable HTTP transport with automatic tool discovery. Turn your Python CLI tools into MCP servers Claude Code can call directly.
Audit a project, plan the path to a win condition, then execute step by step with real-world verification. A fresh-context verifier sub-agent checks the world before any success is declared.
★not rated 5 3mo agoA
tokens not measured
originalMIT
Spec-driven workflow with two independent cross-model review gates, an append-only hardening ledger with an escalation ladder, and repo-enforced quality. Requires the superpowers plugin.
★not rated 5 2d agoA
tokens not measured
originalMIT
Maestro Frontier is an opt-in, zero-dependency local multi-CLI fusion engine for AI coding agents. Use /maestro:frontier catalog to inspect Opus 5 and current Codex models, then compose a panel with validated effort. Panel, judge, and synthesizer subprocesses are one-shot and read-only. Optional policy hooks are…
★not rated 5 yesterdayA
tokens not measured
originalMIT
A complete session management system for Claude Code. Memory persistence, session handoffs, context awareness, and delegation basics - extracted from a 650-node production AI system.
★not rated 5 20d agoA
tokens not measured
originalMIT