Featured on Orynth AIKnowledgeCMSAGENT LOOP CMS
Loop Engineering for Website Growth

A CMS operated by
an agent loop, not by you.

An AI agent runs an endless growth loop over your knowledge site — sensing its health, researching what matters, triaging issues, acting, creating and distributing content, then feeding every result back into the next cycle. There is no admin panel. There is a loop.

View on GitHub See the Loop ↓
⚙ Phase 4 — MEASURE is live: per-article search metrics, page-2 opportunity mining, gated rewrites. Dashboard →
The Growth Loop

One tick, six stages

Each cycle is a resumable "tick" — not a runaway while(true). A scheduler fires the tick; the agent walks the stages guided by playbooks; every result is measured and written back.

1SENSE

Monitor the site

Analytics, Search Console, access logs, revenue and uptime become structured observations in the loop's memory.

2RESEARCH

Collect information

Trends, sources and competitor signals are gathered by collector adapters — the raw material for decisions.

3TRIAGE

Organize the issues

Observations become a prioritized issue queue: traffic drops, indexing errors, content gaps, stale pages.

4ACT

Fix what's broken

The agent works issues via playbooks — SEO fixes and site changes land through git worktrees as pull requests.

5CREATE

Generate content

Articles, videos and share cards are generated — and must pass quality gates (a separate verifier agent) before anything ships.

6DISTRIBUTE

Publish & announce

Publisher and announcer adapters push content to the site, social feeds and video platforms.

MEASURE — results flow back into SENSE. The loop learns what grew the site, and does more of that.
Architecture

Six components, agent-agnostic

The agent is replaceable — any agent CLI (claude -p, codex exec) can execute a stage. The loop is the product.

Ticks & scheduling

cron/systemd fires resumable ticks per cadence — sense hourly, create daily, act weekly. Crash-safe: every tick starts from persisted state.

📖

Playbooks

Procedures are first-class files. The agent must read the playbook before executing a stage — no improvisation on production.

🔌

Adapters

Pluggable sensors / collectors / generators / publishers / announcers. Swap Search Console for Plausible, FTP for GitHub Pages.

🛡

Quality gates

Creator and verifier are separate agents. Content that fails factuality or SEO gates is repaired or discarded — never silently published.

🗄

Memory

SQLite holds the issue queue, content ledger, metrics time series and a decision log you can audit.

🌿

Worktrees

Site-modifying actions run in isolated git worktrees and land as reviewable pull requests.

loopfile.yaml
site: aiknowledgecms.exbridge.jp
kpi: [pv_weekly, affiliate_clicks, indexed_pages]
cadence:
  sense: hourly
  create: daily
  act: weekly
budgets:
  articles_per_day: 3
  llm_cost_per_tick_usd: 2.0
gates: [factuality_check, seo_check]
escalate_when: [revenue_drop_30pct, index_errors]
adapters:
  sensors: [gsc, simpletrack]
  publisher: ftp_heteml
  announcer: aixsns
Declarative

One Loopfile per site

The whole loop — KPIs, cadence, budgets, gates, escalation — is declared in a single file. Point the runner at a Loopfile and the site starts growing.

  • KPIs define what "growth" means for this site — the loop optimizes for them
  • Cadence controls how often each stage runs, independently
  • Budgets and gates are enforced by the runner, not left to the agent's judgment
Loop Engineering

Loops amplify judgment —
so brakes are part of the spec

An autonomous infinite loop without guardrails is a liability. These are built in, not bolted on.

💰

Budgets

Hard caps per tick: LLM cost, article count, API calls. The loop stops before your wallet does.

🧪

Dry-run

Every stage can run with full reasoning and zero side effects. Audit before you trust.

🛑

Kill switch

One flag halts the loop. State is persisted, so resuming is safe.

🙋

Escalation

Declared conditions page a human (email / LINE). Verification stays a human responsibility.

Dogfooding

This site is the reference deployment

aiknowledgecms.exbridge.jp itself is being rebuilt as the framework's first instance: an AI-agent-economy media site grown by the loop — articles, videos, share cards and revenue, all produced and measured by the loop, with every tick's report published. The framework's proof is the site you are reading.

Roadmap

Where this is going

P0

Concept & spec ✓

Manifesto, Loopfile specification, architecture. Public from day one.

P1

Core tick runner + SENSE / TRIAGE / REPORT ✓

Real sensors populate the issue queue in SQLite. Every tick publishes a public loop report — running hourly on this site.

P2

CREATE / DISTRIBUTE + quality gates ✓

The loop researches sources, generates articles, and publishes only what passes an independent verifier gate — rejected drafts stay in the ledger with reasons.

P3

ACT + share featuresIN PROGRESS

Live dashboard shipped — the loop publishes its own scoreboard hourly. Next: growth cards, a diagnosis experience, and site-modifying playbooks via worktrees.