Native Claude-Code agentic loop: brainstorm → plan → execute → review → e2e → pr → demo (+ ad-hoc fix). Auto-routes single vs parallel-wave mode, with a hardened environment probe driving capability-gated in-loop E2E and a stall watchdog that recovers hung/silently-killed agents and reports periodic progress. Replaces the Docker-hosted ralphex flow and the retired dev-workflow plugin. Includes auto-generated narrated walkthrough demos (TTS → paced capture → mux → S3).
Plugin ralph from marketplace georg-malahov-claude-skills in georg-malahov/claude-skills. Commands read from the repo's manifest — 22 days ago.
/ralph:brainstormRalph brainstorm — interview-driven context gathering, output feeds /ralph plan.
/ralph:demoAuto-generate a narrated walkthrough video (screenplay → TTS → paced capture → mux), host on S3, attach to the PR.
/ralph:e2eRalph e2e — consume FIXME(e2e) placeholders and implement E2E tests one at a time.
/ralph:executeRalph execute — native ralph-loop. Auto-detects single-plan vs wave mode.
/ralph:fixRalph fix — turn multiple bug reports/observations into committed test-first fixes.
/ralph:planRalph plan — grill + auto-route single vs parallel; emit plan(s) and (if parallel) manifest.
/ralph:prRalph PR — create a PR with selectable hardening level (lean vs hardened).
/ralph:ralphRalph — native agentic loop. Dispatches to brainstorm | plan | execute | review | e2e | pr | demo | fix.
/ralph:reviewRalph review — browser-driven visual walkthrough. The agent verifies in the running app BEFORE concluding; the human is consulted only where judgement is needed.
ralph-documentationDocumentation reviewer — flags README, ADR, and inline-doc updates needed for the current changes. Used in ralph review loop iteration 1 only. Read-only.
ralph-fixerReceives the full findings list from one round of ralph review-agent fan-out and applies fixes. Decides what's real and actionable. Commits each fix with a clear message. Re-runs lean validation and fixes any pre-existing bugs that surface. Outputs a FIXES section listing fixed vs skipped (with reas
ralph-implementationImplementation-correctness reviewer — verifies code meets the plan's stated goals. Used in the ralph review loop on both comprehensive (iter 1) and critical (iter 2-5) passes. Read-only.
ralph-qualitySecurity, multi-tenant safety, and TS-correctness reviewer for T3 + ZenStack + Better Auth apps. Use during the ralph review loop (iteration 1 comprehensive + iterations 2-5 critical-only re-check). Read-only.
ralph-simplificationOver-engineering and duplication reviewer — flags dead code, premature abstractions, redundant indirection. Used in ralph review loop iteration 1 only. Read-only.
ralph-taskImplements ONE task from a ralph plan. Receives the plan file, the single task description, and the task contract from prompts/task.md. Edits code, runs lean validation (lint + typecheck + test:unit), commits on green. Never runs E2E. See prompts/task.md for the full contract.
ralph-testingUnit-test coverage and quality reviewer for Vitest projects. Flags missing unit coverage, brittle assertions, untested branches. Does NOT review E2E — E2E lives in /ralph e2e. Used in ralph review loop iteration 1 only. Read-only.
Turn photos of paper documents into scanner-grade A4 PDFs: perspective corrected, even white paper, neutral ink with coloured signatures and stamps preserved, desk and binding removed, exact A4 page. Several photos combine into one multi-page PDF. CLI, HTTP service and a container.
1 skill
Process videos: optimize for web, transcribe with Deepgram Nova 3 or a cheaper Gemini analysis engine (audio transcript-cheap + dev-video screen analysis with auto screenshots), add subtitles in any language, translate, and burn into video with customizable styles.
1 skill
Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Talk like caveman. Cut filler. Keep technical details explicit.
22 skills · 1 command · 3 agents · hooks
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.
1 skill
Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications
5 skills · hooks
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
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.