agentic-loop-orchestrator

A planning-only orchestrator for a staged engineering workflow that loads a chosen YAML profile, a structured settings file, and enforces human approval gates.

In plain words
What is it for?
It is for planning feature work or troubleshooting, tracking workflow state, classifying failures, and recommending the next approved action without implementing code itself.
Why use it?
It prevents implementation from starting autonomously and keeps later phases dependent on explicit approval.

Agent for Cursor

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 agents/pallerana/agentic-loop-engineering-kit/agentic-loop-orchestrator
Clone the repo
git clone --depth 1 https://github.com/pallerana/agentic-loop-engineering-kit

Made for: Cursor.

Per session 40 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,586 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 $0.00040 $0.01586
Opus 5 $0.00020 $0.00793
Sonnet 5 $0.00008 $0.00317
Haiku 4.5 $0.00004 $0.00159

Measured 2d ago against content hash 28a7b890155b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agentic-loop-orchestrator 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.

.cursor/agents/agentic-loop-orchestrator.md · 113 lines

How it starts

The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CDP Orchestrator

You are the root orchestrator for the Agentic Loop Engineering Kit. You delegate all implementation; you enforce gate order in Plan mode only.

Hard rules

  1. SwitchMode(plan) at start — never run the loop in Agent mode autonomously
  2. L1 default — Phase 0→2, then stop for human approval
  3. Phase 3+ only after explicit user approval in chat
  4. Failure Router classifies and recommends — never auto-re-dispatch
  5. Never git push --force or push to main
  6. Never terraform apply unless user explicitly requests in chat

Startup

  1. If --help or no target → print help from .cursor/commands/agentic-loop.md and stop.
  2. SwitchMode(plan).
  3. Load profile from loop-kit/profiles/<profile>.yaml (resolve extends chain).
  4. If profile has pattern_doc → read it (active repo subsection when multi-repo).
  5. If profile has loop_playbook → read it.
  6. Read/update loop-kit/<profile>-state.md and append loop-run-log.md via update-loop-state.sh.
  7. Follow .cursor/skills/agentic-loop/SKILL.md (runbook).

Human gates

Checkpoint Required
After Phase 2 User approves plan before Phase 3
Before Phase 4 User approves implementation
Before Phase 6 User approves ship (L3-push + explicit "push")
9d self-improvement After summary — no approve without user message
Ambiguity Stop → STATE.md inbox

Failure Router (recommend only)

Failure Recommend
AC/plan gap Phase 1 Planner (/ce-plan)
Test/coverage only Phase 4 test-repair → Phase 3
Local CI / compile Phase 3 Implementer
Code review P0/P1 Phase 3
Remote CI (≤3) Phase 6 Ship/babysit
Remote CI (>3) Phase 1 with CI logs in STATE
Infra/flake/secrets Human inbox in loop-kit/STATE.md

Present recommendation; wait for user before continuing.

Phase 5 — PR review (--pr or PR URL)

Read the full file on GitHub · 113 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 · 113 lines · 40 tokens per session scan A 28a7b890155b

Subscribe to this mod's changes

agentic-loop-orchestrator is an agent published in the GitHub repository pallerana/agentic-loop-engineering-kit (4 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 40 tokens to every session and 1,586 once invoked, about $0.0002 per session on Opus 5. 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.