skills.rest
lightrain-a
lightrain-aCOMMUNITY
Explore all skills by this author →
GitHub

neat-freak

Start ARIS autonomous rounds from compact live-state artifacts and reconcile TODO and job manifests.

Updated May 20, 2026
Search Tags:#state synchronization#aris#execution orchestration#remote gpu jobs#autonomous rounds#progress reconciliation#document hygiene
One-click install
npx skills add https://github.com/lightrain-a/medtrace-aris --skill neat-freak-lightrain-a
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: neat-freak
Source: https://github.com/lightrain-a/medtrace-aris/tree/main/.codex/skills/neat-freak
Command: npx skills add https://github.com/lightrain-a/medtrace-aris --skill neat-freak-lightrain-a

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents ARIS from repeatedly rereading long research artifacts by re-centering execution on compact, current state files while keeping task execution synchronized across rounds.

Core Features & Use Cases

  • Fast-Start Read Order: Prioritizes execution snapshot, TODO queue, remote GPU job manifest, supervisor state, and referenced progress files to answer the next action quickly.
  • On-Demand Long Reads: Loads large documents only when the next action requires clarification on scope, method/claim logic, literature/novelty changes, or paper-facing edits.
  • State Contract & Round Hygiene: Defines clear “single source of truth” files and update requirements for snapshot, TODO, status logs, and remote job provenance after substantive rounds.
  • Job Running Guardrails: Enforces evidence-based “running” detection, prevents duplicate launches into shared output directories, and marks stale-progress jobs as BLOCKED/VERIFY.
  • Use Case: When an ARIS executor is continuing research work or handing off between autonomous rounds, it uses this skill to reconcile status/TODO, update provenance, and resume with minimal I/O.

Quick Start

Use the neat-freak skill to resume the next ARIS autonomous round by reading the execution snapshot and TODO queue first, then executing the highest-priority concrete task and updating the snapshot, TODO, status, and remote job manifest.

Frequently Asked Questions about neat-freak

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I start autonomous rounds from live state without rereading long research documents?▼

Start autonomous rounds from live state by reading the execution snapshot and TODO queue first, loading large documents only when scope clarification or paper-facing edits are needed. This prioritizes compact state files to answer the next action quickly.

What is execution state synchronization for remote GPU jobs in ongoing experiments?▼

Execution state synchronization tracks remote GPU jobs using a defined read order across execution snapshots, TODO queues, and job manifests. It enforces evidence-based running detection and prevents duplicate launches into shared output directories.

How do I reconcile progress and status logs between autonomous research rounds?▼

Reconcile progress between autonomous rounds by applying strict update rules to snapshot, TODO, status logs, and remote job provenance after substantive work. This ensures the next round resumes with minimal I/O and accurate state.

How do I prevent duplicate launches into shared output directories for remote GPU jobs?▼

Prevent duplicate launches by enforcing job running guardrails that require evidence-based running detection. The skill checks the remote GPU job manifest and marks stale-progress jobs as BLOCKED or VERIFY before authorizing new launches.

When do I need to load full research documents instead of compact state files?▼

Load full research documents on demand when the next action requires clarification on scope, method or claim logic, literature and novelty changes, or paper-facing edits. Compact state files handle routine execution and status updates without long reads.

Can I use compact state files for experiment handoff cleanup and status reconciliation?▼

Use compact state files for experiment handoff cleanup and status reconciliation by defining a single source of truth for snapshots, TODO queues, and status logs. Update provenance after substantive rounds to maintain document hygiene.

Related Skills

Explore More Software Engineering Skills→
NVIDIA-TAONVIDIA-TAO
83

tao-run-on-slurm

Automate remote TAO GPU job submission on SLURM clusters via SSH.

Official
Advanced
s977043s977043
2

working-context

Load and update local worktree ticket context for Cloud task handoffs.

Community
Intermediate
bretbouchardbretbouchard
2

workflow-reconciliation

Reconcile workflow state across sessions to prevent checkpoint drift.

Community
Advanced
LingrongyeLingrongye
1

run-experiment

Deploy and manage ML experiments on local or remote GPU resources.

Community
Advanced