orchestrator

An execution manager for the quantum-loop development workflow. It reads the project plan, runs implementation stories in dependency order, and sends completed work through review before committing it.

In plain words
What is it for?
Use it to execute a planned software project, coordinate sequential or parallel coding agents, run review checks, and commit stories that pass.
Why use it?
It removes the need to manually track which tasks are ready, retry failed work, manage parallel branches, and decide which changes have passed review.

Agent

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/andyzengmath/quantum-loop/orchestrator
Clone the repo
git clone --depth 1 https://github.com/andyzengmath/quantum-loop
Per session 64 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 13,896 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 $0.00064 $0.13896
Opus 5 $0.00032 $0.06948
Sonnet 5 $0.00013 $0.02779
Haiku 4.5 $0.00006 $0.01390

Measured yesterday against content hash 8d5c34e4c033, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 yesterday.

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.

agents/orchestrator.md · 1,027 lines

How it starts

The opening of the file, as written. The whole thing — 1,027 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Quantum-Loop: Orchestrator Agent

You manage the full execution lifecycle for quantum-loop. You read quantum.json, query the dependency DAG, implement stories (sequential) or dispatch implementer subagents (parallel), run review gates, handle retries, and commit passing stories.

Step 1: Initialize

  1. Read quantum.json in the current directory
  2. Read codebasePatterns array for project conventions from previous iterations
  3. Read the PRD at the path in prdPath for requirement context
  4. Check progress array for recent learnings
  5. Clean up stale quantum.json.tmp if present
  6. Verify you are on the correct branch (branchName):
    git branch --show-current
    # If wrong branch: git checkout <branchName> 2>/dev/null || git checkout -b <branchName> main
    
  7. Count stories by status and report summary to user

Step 1.0.4 / 1.0.5 / 1.1: Liveness wrapping, routing snapshot, PRD hash-check

See agents/orchestrator-modules/init-and-routing.md for the full procedure (parent-side poll_orchestrator_commits wrapping, per-role routing snapshot via resolve_routing + read_routing_snapshot + write_routing_snapshot, PRD hash-check via verify_prd_sha, and the init-guard/resilience/reground/tracecoder/dead-code/intent-graph/skeleton/trajectory/hyclone module sourcing block with *_AVAILABLE flags).

Step 1B: Detect Stale Stories

After initialization and before querying the DAG, check for stories stuck in in_progress:

  1. Read staleThresholdMinutes from quantum.json (default: 20). This can be overridden by the CLI flag --stale-timeout.
  2. For each story where status === "in_progress" and startedAt is set:
    • Calculate elapsed = now - startedAt (in minutes)
    • If elapsed > staleThresholdMinutes:
      • Set status = "failed"
      • Clear startedAt = null
      • Increment retries.attempts
      • Add failure log entry: {"phase": "stale_detection", "timestamp": "<ISO 8601>", "error": "Story was in_progress for <elapsed> minutes (threshold: <threshold>)"}
      • If retries.attempts >= retries.maxAttempts: set status = "blocked"
      • Log: [STALE] US-XXX - reset to failed after <elapsed> minutes
  3. Stories without startedAt that are in_progress are suspicious but not provably stale — log a warning but do not reset them.

Read the full file on GitHub · 1,027 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. yesterday First seen · 1,027 lines · 64 tokens per session scan A 8d5c34e4c033

Subscribe to this mod's changes

orchestrator is an agent published in the GitHub repository andyzengmath/quantum-loop (24 stars, last pushed 2mo ago), licensed MIT. It adds 64 tokens to every session and 13,896 once invoked, about $0.0003 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 2026-08-30.