stage-snapshot-writer

Capture operator-visible stage outputs into task-scoped snapshot files with atomic writes.

11|1|Updated Apr 10, 2026
One-click install
npx skills add https://github.com/Arcanada-one/datarim --skill stage-snapshot-writer
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: stage-snapshot-writer
Source: https://github.com/Arcanada-one/datarim/tree/main/skills/stage-snapshot-writer
Command: npx skills add https://github.com/Arcanada-one/datarim --skill stage-snapshot-writer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

It prevents lost context by reliably writing each stage’s final, operator-visible snapshot so later commands can resume with consistent state.

Core Features & Use Cases

  • Deterministic stage snapshotting: Records the Summary + Gate Results + CTA block emitted by each /dr-* command into a task-scoped snapshot file.
  • Overwrite semantics for replays: Replaces an existing snapshot for the same task/stage to keep results current.
  • Safe, bounded outputs: Enforces an 8 KB total size cap with a truncation marker to avoid runaway context growth.
  • Concurrency control: Uses a per-task lock to prevent concurrent writers from corrupting snapshot contents.
  • Kill switch: Can disable writing entirely via DATARIM_DISABLE_SNAPSHOT=1.

Quick Start

Use stage-snapshot-writer to write the snapshot for task TASK-ID after you generate the final CTA block in stage plan using the /dr-plan command.

Frequently Asked Questions about stage-snapshot-writer

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

FAQPage Schema
How do I persist AI agent pipeline stage outputs for context recovery?▼

You can persist AI agent pipeline stage outputs by capturing the final Summary, Gate Results, and CTA block into a task-scoped snapshot file for later replay and consistent state recovery.

What is the best way to prevent lost context after terminal shutdown in workflow orchestration?▼

Preventing lost context after terminal shutdown is achieved by reliably writing each stage's operator-visible snapshot file, allowing later commands to resume with consistent state.

How do I handle concurrency locking when multiple agents write stage snapshots?▼

Concurrency locking for multiple writers is handled using a per-task lock mechanism that prevents concurrent writers from corrupting snapshot contents during atomic write and rename operations.

Does this stage snapshotting approach enforce size limits on context persistence files?▼

Yes, this stage snapshotting approach enforces safe bounded outputs with an 8 KB total size cap, applying a truncation marker to avoid runaway context growth in the snapshot file.

Can I disable context persistence snapshots entirely in my CLI workflows?▼

Yes, you can disable context persistence snapshot writing entirely by setting the optional no-op kill switch environment variable DATARIM_DISABLE_SNAPSHOT to 1.

Do I need specific CLI commands to trigger stage snapshotting in AI pipelines?▼

You need to use producer-style `/dr-*` flows across plan, prd, do, qa, verify, and auto stages to generate the final CTA block that triggers the snapshot writing.