ceo-loop

An automated decision-making loop for project governance. It repeatedly reviews pending decisions and project state, uses earlier decisions as context, makes or escalates decisions, and records the results.

In plain words
What is it for?
Use it to process pending project decisions, approve or modify them, escalate uncertain cases, and keep a journal of decisions and loop progress.
Why use it?
It provides a repeatable way to handle project decisions over multiple cycles instead of leaving them scattered or unattended.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/nxtg-ai/forge-plugin/ceo-loop
Clone the repo
git clone --depth 1 https://github.com/nxtg-ai/forge-plugin
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,331 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.01331
Opus 5 $0.00000 $0.00665
Sonnet 5 $0.00000 $0.00266
Haiku 4.5 $0.00000 $0.00133

Measured 2d ago against content hash 14c7a8c32c6d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ceo-loop scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

docs/commands/ceo-loop.md · 108 lines

How it starts

The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/forge:ceo-loop

Activate the CEO Decision Loop in ORBIT mode -- a continuous multi-iteration governance cycle that observes pending decisions, reasons with precedent, builds decisions, inspects via retrograde, and turns for the next iteration.

Level L1 Vibe Coder
Category CEO Decision Loop
Syntax /forge:ceo-loop [max-iterations] [time-limit-minutes]

What It Does

/forge:ceo-loop starts an autonomous decision-making engine that runs in a continuous ORBIT cycle: OBSERVE pending decisions and project state, REASON about each decision using precedent from the decision journal, BUILD the actual decisions (approve, modify, escalate), INSPECT via retrograde analysis of previous decisions to calibrate accuracy, and TURN to prepare for the next iteration. The loop persists across Claude Code's Stop events via a hook (ceo-loop-stop.sh) that automatically triggers the next iteration.

The loop manages its own state: .claude/ceo-loop-state.json tracks iteration count, trust level, accuracy metrics, and timing. .claude/ceo-loop-decisions.jsonl is an append-only journal of every decision made. .claude/ceo-loop-progress.md bridges context across compaction events. The trust system starts at "standard" and can escalate based on decision accuracy -- more correct decisions earn higher trust, which unlocks deeper autonomous authority.

Without this command, governance decisions accumulate in .claude/ceo-decisions-pending.json waiting for manual review. The CEO loop processes them systematically, with appropriate depth for each decision type (trivial decisions are auto-approved, complex decisions get deep analysis, risky decisions are escalated to the human).

Syntax & Options

/forge:ceo-loop [max-iterations] [time-limit-minutes]
Option Description
max-iterations Maximum number of ORBIT iterations to run (default: 20, range: 1-50)
time-limit-minutes Maximum wall-clock time before the loop stops (default: 30, range: 5-120)

Read the full file on GitHub · 108 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 108 lines · 0 tokens per session scan A 14c7a8c32c6d

Subscribe to this mod's changes

ceo-loop is a command published in the GitHub repository nxtg-ai/forge-plugin (5 stars, last pushed 13d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,331 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.