jobs

Name: jobs
Availability: InStock
Author: kamilseghrouchni

Coordinate long-running batch jobs with a persistent ledger and parallel leaf-level tasks.

Updated Mar 31, 2026

Search Tags:#orchestration #long-running #context-management #parallel-processing #ledger #batch-jobs #cli-tools

Ready to run

npx skills add https://github.com/kamilseghrouchni/vcro-sourcing --skill jobs-kamilseghrouchni

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: jobs
Source: https://github.com/kamilseghrouchni/vcro-sourcing/tree/main/.claude/skills/jobs
Command: npx skills add https://github.com/kamilseghrouchni/vcro-sourcing --skill jobs-kamilseghrouchni

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Running long-running batch jobs in vCRO often exceeds memory in a single conversation. This skill externalizes working memory to a persistent ledger, allows parallel task execution at the leaves, and preserves progress so runs survive context compaction and remain inspectable.

Core Features & Use Cases

Externalize working memory to store/runs/<slug>/ledger.md to track task status across waves.
Parallelize at the leaves to maximize throughput while avoiding memory bottlenecks in the orchestrator.
Maintain a digest-based Task Ledger and a Verification Log to provide traceability and provenance.
Pre-flight cost/time estimates and explicit user confirmation before spawning waves.
Caching and safeguards to prevent re-ingesting papers and to manage noisy inputs.

Quick Start

Initialize a new long-running job by creating the ledger under store/runs and launching parallel waves after the pre-flight approval.

Frequently Asked Questions about jobs

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

FAQPage Schema

How do I run long-running batch jobs without exceeding context memory limits?▼

Long-running batch jobs avoid exceeding memory by externalizing working memory to a persistent ledger at store/runs, preserving progress so runs survive context compaction and remain inspectable.

What is the best way to parallelize data workflows over large document sets?▼

Parallelizing data workflows over large document sets is best handled by fanning out tasks at the leaves to maximize throughput while avoiding memory bottlenecks in the orchestrator.

How do I maintain task traceability and provenance during parallel processing?▼

Maintaining task traceability during parallel processing requires a digest-based Task Ledger and a Verification Log to track task status across waves and provide provenance.

Can I estimate batch processing costs and time before starting a data workflow?▼

Batch processing costs and time can be estimated through pre-flight estimates with explicit user confirmation required before spawning parallel waves of tasks.

How does caching prevent re-ingesting documents in long-running orchestration workflows?▼

Caching in long-running orchestration workflows prevents re-ingesting documents and manages noisy inputs through built-in safeguards designed to optimize large document set processing.

When should I use a persistent ledger for context management in batch jobs?▼

A persistent ledger for context management in batch jobs should be used when workflows exceed memory, require persistence across context compaction, and involve tasks that can be parallelized over leaf-level agents.