resource-discovery
Resource Discovery Loop
Sweep committed sources and rotating queries each week, score candidates against the resource selection notes, and propose ready-to-merge entries for the curated library without publishing them.
Loopify first
Paste this into your agent
Turn a manual workflow into a draft loop bundle. The skill asks questions before it writes files.
Fetch https://loopmaster-ai.pages.dev/skills/loopify/SKILL.md and follow it.
Start with the Q&A intake and ask only for missing details.
Do not schedule, publish, spend, request credential values, or make destructive changes until I approve.
Workflow: <describe the recurring job you want turned into a loop>For people who code with AI
Turn work you repeat by hand into a loop your agent runs.
A loop is an agent doing a recurring job — on a trigger, with checks that prove it worked, and a rule to stop and ask you before anything risky. Loopmaster gives you tested templates and a copy-paste start. It runs on its own loops, so you can see the whole thing working.
Start here
Paste this into your agent
Works with Claude, Codex, Hermes, or any agent that can fetch a URL. Edit the one line in <angle brackets>.
Read https://loopmaster-ai.pages.dev/llms.txt — it explains how this site works
for agents.
I have a recurring job I want to turn into an AI loop:
<describe it — e.g. "every Monday, check our docs for dead links and
open a PR fixing them">
Use Loopmaster's quickstart and skills index to pick the best starting
template or skill. Before you install anything, tell me:
1. what you'll install and where it will run,
2. which checks will prove it's working,
3. what needs my approval — anything that runs on a schedule, spends
money, publishes publicly, or needs new credentials.
Nothing goes on a schedule and nothing gets published until I approve it.Get started
Paste one prompt into your agent and ask it to choose, install, verify, and report on a loop before anything runs unattended.
Learn
Short lessons teach the five-beat loop shape, trustworthy anatomy, checks, guardrails, harnesses, and maintenance.
Browse templates
Pick from loop templates with a guide for you, an install doc for your agent, and an honest status: tested on a date you can see, or 'not tested yet'.
What is a loop?
An AI loop is an agent running tools in a loop to achieve a goal. A model call is a part. A loop is the working system.
TRIGGER → CONTEXT → ACT → VERIFY → escalate?
what what the do the check repeat,
starts agent work against stop, or
it knows criteria ask a human- Trigger — what starts it: a schedule, a failure, a goal.
- Context — what the agent knows: docs, state, memory.
- Act — what it does: call tools, edit files, run commands.
- Verify — how it checks the result: tests, sources, diffs.
- Escalate — what it does when stuck: repeat, stop, or ask a human.
This site maintains itself
Loopmaster runs on the loops it publishes. Loops propose; checks run; people approve anything public-facing.
link-check
Link & Source Check Loop
Check Loopmaster documentation and resource URLs, update safe maintenance stamps, and escalate source updates.
content-freshness
Content Freshness Loop
Check repository content for stale pages, dead links, superseded claims, and missing ownership records, then propose source-backed updates without publishing rewrites automatically.
Learn loop engineering
8 short lessons, from “what is a loop?” to running one you trust unattended.
Lesson 1
What is an AI loop?
The 30-second and the five-minute answer to what an AI loop is, and why the loop — not the model call — is the unit of value.
Lesson 2
Loop patterns and when to use them
How to match a goal to a loop pattern, when not to automate at all, and why every pattern carries a characteristic failure you sign up for.
Lesson 3
Anatomy of a trustworthy loop
The structural checklist a loop must declare before anyone should run it unattended — the prose twin of the loop.yaml spec.
Lesson 4
Verification: the actual product
Anyone can schedule an agent; the hard part is trusting the result. This module documents, honestly, what this site's own 'verified' status means.
Lesson 5
Guardrails and human gates
How much autonomy to grant a loop, and the short list of things that always need a human. Carries the site's trust posture; no autonomy hype.
Lesson 6
Harness engineering
The model is not the agent; the harness — tools, prompts, memory, hooks, guardrails — is what does the work.
Lesson 7
Loop maintenance and rot
Loops decay silently; maintenance is a design requirement, not ops. This site's own roster is the worked example.
Lesson 8
Your first loop in 15 minutes
Go from zero to a running, verified research-watch loop. References the template and skill pack; does not duplicate their commands.
The reading room
The best writing on agents and loops, selected for practical value and kept current by this site's discovery and link-check loops.
Skills your agent can copy
Installable packs that help an agent choose a template, run checks, and stop at your approval gates.
authoring
Loopify
Use when a human describes a manual workflow and needs Q&A intake plus a draft Loopmaster loop bundle: loop.yaml, README, AGENT-INSTALL.md, verify.sh, and catalog-entry instructions.
authoring
Deploy Your First Loop
Use when an agent or human wants to deploy the research-watch Loopmaster template through cron, GitHub Actions, or agent-kanban and verify a running loop in one session.
Have a workflow you run by hand? Loopify turns it into a draft template.
6tested templates
3not tested yet
40curated resources
2agent skills
For agents
This site is machine-readable
llms.txt · registry/index.json · skills/index.json · spec · agent-guide
Start at llms.txt — it explains how this site works for agents. Three fetches reach any template manifest.