ralph
claudeRalph Loop — an autonomous agentic coding workflow with sandbox isolation, scoped memory, and self-managing task iteration.
4 skills · 4 agents · 1 MCP server · hooks
A spec-driven development workflow that turns a rough idea into a verified implementation. It elicits and hardens an acceptance-criteria-first spec against your codebase, compiles it into an execution driver, then checks the built code against every criterion — with read-only adversarial judges gating quality at each stage and every spec change committed automatically. When OpenAI Codex CLI is available it adds an optional, fail-open second judge of a different provider, so every quality gate is cross-checked rather than one model auditing itself.
Plugin spec-ops from marketplace zaksak in zakattack9/agentic-coding. Commands read from the repo's manifest — 4 days ago.
launch-specCompile a finished, verified spec into the self-contained /goal driver prompt that implements it — then stop. Use this on a finished, refined spec to turn it into the one-shot execution prompt you'd otherwise hand-type: the goal, the spec + checklist references, the boundaries (what NOT to touch), a
loop-specCompile a spec into a self-contained review→fix CONVERGENCE-LOOP driver prompt that a fresh session runs to harden the spec until an independent sweep finds nothing material — then stop. Use it for a heavy, unbiased, cross-model spec-hardening pass in a dedicated session: "loop on this spec", "conve
orchestrate-specRun the whole spec workflow — write → refine → launch → build → verify — end-to-end in ONE session, so you never hand-run each spec-ops skill yourself. Use when the user says "take this idea all the way to a verified implementation", "orchestrate the spec", "drive the whole spec pipeline", "go from
refine-specReview, verify, and tighten an existing feature spec until it is accurate, hallucination-free, not over-engineered, and ready to implement. Use immediately after write-spec, or whenever the user asks to review, fact-check, refine, simplify, de-risk, or finalize a spec / PRD / requirements doc before
verify-specVerify that an implementation, change, or audit result actually matches the claims made about it — grounded against the real codebase, git history, and live read-only system state, never against the spec or any doc. It does NOT check the spec document (that is refine-spec's job); it checks whether r
write-specThe entrypoint to the spec workflow — turn an idea, even a rough one-liner, into a concise, scannable feature spec that contains everything needed to implement the change and nothing more. It first runs a discovery pass (eliciting and distilling requirements via questions, scaled to the change) befo
codex-reviewCross-model Codex reviewer for the loop-spec convergence loop, dispatched as a subagent so the loop can fire Codex CONCURRENTLY — in the same message as its Claude lens reviewers — instead of running a Codex pass afterward. Self-contained: it calls spec-ops's OWN codex bridge directly (no dependency
spec-refine-judgeInternal adversarial readiness judge for the spec-ops refine-spec skill — dispatched by refine-spec, not for general use. Independently reviews a feature spec against the codebase for any remaining inaccuracy, ambiguity, over-engineering, bloat, missing detail, or uncaptured functional / non-functio
spec-verify-judgeInternal adversarial completeness judge for the spec-ops verify-spec skill — dispatched by verify-spec, not for general use. Independently re-derives a verification target's checkable claims and audits a verification ledger against reality, flagging missed claims, weak / standard-mismatched / existe
Ralph Loop — an autonomous agentic coding workflow with sandbox isolation, scoped memory, and self-managing task iteration.
4 skills · 4 agents · 1 MCP server · hooks
Run a small team's entire software lifecycle on GitHub Projects v2 — deterministic, free, GitHub-native, and with no metered AI. It provisions a golden-template board and repo automation across an org, then drives work from issue intake through sprint planning and the full branch → PR → merge flow with check-gated, never-squashed merges, kept safe by GitHub App-token writes and skill-scoped guards.
8 skills
Self-hosting toolkit for managing this skills marketplace: setup-project, promote-skill, sync-skill, new-skill, new-plugin, list-skills, skills-doctor.
7 skills
Manage your Claude Code skills and plugins from inside any project — switch which plugins are active, publish and sync local skills to a central marketplace, and diagnose and auto-repair a broken setup. Adapts to any repo layout and delegates skill authoring to the native skill-creator.
5 skills
Run the OpenAI Codex CLI from inside Claude Code as a cross-model escape hatch: hand a one-off review, question, or write task to a different provider's model for a second opinion, grounded in your repo with web search always on. Every call routes through one fail-open bridge — if Codex is absent, unauthenticated, or broken the skill reports and stops rather than answering as Claude, and Codex's output is surfaced verbatim as untrusted external text. Assumes an OpenAI-authenticated Codex setup.
3 skills · 1 agent
A git worktree workflow for running multiple Claude Code sessions in parallel — no restarts, no manual checkouts. Built on native worktrees, it covers the full lifecycle from spinning up an isolated worktree to finishing it through a guarded, never-squashed PR merge, with a live dashboard of in-flight work and deterministic guards against squash merges and committing unresolved conflicts.
4 skills
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.