Loop Engineering
Design and review autonomous or semi-autonomous agent loops — systems where an agent prompts itself, runs on a schedule or event, and iterates toward a goal without a human at the keyboard. Use this whenever the user wants to build, design, harden, or audit any kind of self-running agent, agentic workflow, control loop, background worker, cron/heartbeat/webhook-driven agent, "Ralph loop", auto-PR-fixer, monitoring agent, or any long-running task that should keep going on its own until done. Trigger this even when the user says "agentic workflow", "automation that runs itself", "an agent that k
Install
npx skills add maxmilian/loop-engineering --skill loop-engineeringgit clone --depth 1 https://github.com/maxmilian/loop-engineering.git /tmp/loop-engineering
mkdir -p ~/.claude/skills
cp -r /tmp/loop-engineering/. ~/.claude/skills/loop-engineeringPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Loop Engineering Loop engineering is the discipline that comes after prompt engineering. A prompt optimizes a single interaction. A **loop** optimizes the autonomous behavior that surrounds it — *when* the agent runs, *what* triggers it, *how* it verifies its own work, *when* it stops, and *when* it hands back to a human. The hard part is never the loop itself. A bare agent loop (call model → run tool → observe → repeat) is ~30 lines and every framework's core is the same. The whole job is the engineering *around* it: verification, stopping conditions, guardrails, context management, and escalation. Build those wrong and you are one edge case away from an agent that loops forever, burns the budget, or silently ships garbage. This skill has two modes. Detect which one the user needs: - **Design mode** — the user is building a new loop. Walk the whiteboard, then the seven principles, then hand them a concrete design. - **Review mode** — the user has a loop already (a script, a workflow, an agent). Use the seven principles as a diagnostic checklist and report what's missing. Most requests are one or the other; some are both. Don't force the framework on a user who just wants a quick answer — but for anything that will run unattended, the failure modes below are real and worth raising. ## Principle 0 — Whiteboard before platform Before writing any code or choosing any framework, answer five questions on paper.
Details
- Skill name
loop-engineering- Repository
- maxmilian/loop-engineering
- License
- MIT
- Repo updated
- 24 June 2026
FAQ
What does the Loop Engineering skill do?
Design and review autonomous or semi-autonomous agent loops — systems where an agent prompts itself, runs on a schedule or event, and iterates toward a goal without a human at the keyboard. Use this whenever the user wants to build, design, harden, or audit any kind of self-running agent, agentic workflow, control loop, background worker, cron/heartbeat/webhook-driven agent, "Ralph loop", auto-PR-fixer, monitoring agent, or any long-running task that should keep going on its own until done. Trigger this even when the user says "agentic workflow", "automation that runs itself", "an agent that k
How do I install the Loop Engineering skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add maxmilian/loop-engineering --skill loop-engineering. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Loop Engineering popular?
It lives in maxmilian/loop-engineering, which has 29 GitHub stars and 2 forks.
Related skills
Reverse Skill Evolver
252233admin/reverse-skill-evolver
Route authorized reverse-engineering, cybersecurity, and CTF tasks through the bundled skills and tool workflows. Use for APK/mobile, native binaries, JavaScript signatures, captured protocols, security assessments, Forefy ASR discovery, and bounded routing evolution.
Lammps Workflows
29chatmaterials/lammps-workflows
Use when the task involves LAMMPS atomistic simulation workflows, including input script creation, minimize, NVT, NPT, and restart setups, force-field and data-file checks, log.lammps review, restart handling, and scheduler scripts.
Analysis to Delivery
26BlueprintOS/analysis-to-delivery
通用需求分析到开发设计工作流 — thin Skill router,加载规则与路径按声明。Use when starting any new feature requiring structured analysis-to-delivery workflow.
Deep Ppt
29jiadizhunine/deepPPT
Use when creating or revising research-style PowerPoint decks for lab meetings, journal clubs, defenses, weekly reports, or conferences, especially Chinese or English academic slides that need a resumable agent workflow, the deepPPT blue title system, evidence-led layouts, selective red key findings, package checks, and rendered visual review.
Executing Hard Tasks
26wsbm393/Fable-Skill
Use when a task spans multiple steps, files, or unknowns — before planning the order of work, whenever about to report completion or success in any wording, and when picking the next action after a result, failure, or surprise. Also under deadline pressure, or when resuming work built on questionable foundations.
Sci Select
30keros68/sci-select
Use when a user needs SCI/SCIE/ESCI/SSCI journal-submission help: evidence-backed candidate-journal discovery from a title, abstract, keywords, or manuscript text; public-metric and risk lookup for a known journal; or a pre-submission compliance check of a draft against a chosen journal's Author Guidelines and same-journal conventions. Also for Chinese requests such as 选刊、投稿期刊推荐、期刊查询、中科院分区、投稿前检查、期刊要求核对. This skill produces candidates, evidence, checks, and uncertainty; it does not judge research quality, predict acceptance, choose a uniquely best journal, draft prose, or simulate peer review.
Browse the most popular Claude skills, all productivity & notes skills, or write your own with the skills generator.