Does the final document contain the tracking markers?

No, the /stream.finalize command performs a final integrity check and then strips all tracking markers, hashes, and YAML metadata to produce a clean, professional markdown file.

What happens if Claude's context limit is hit mid-generation?

The skill detects the incomplete section via the status command, preserves the partial content for context, and allows you to run the resume command to continue exactly from the interruption point.

Why should I use streaming-output instead of standard file appends?

Standard shell appends like heredocs can fail silently or corrupt data when AI context limits are reached mid-write. This skill uses specialized markers and hashes to ensure every section is verified and recoverable.

Can I use custom structures for my documents?

Yes, you can initialize a custom section plan using the /stream.init command with a comma-separated list of IDs, or use pre-defined templates for common formats like B-SPEC.

Streaming Long-Form Content

Name: Streaming Long-Form Content
Author: ddunnock

byddunnock

•

内容管理

Streams long-form documents to markdown files with incremental saving and automatic recovery capabilities.

关于

The streaming-output skill is designed for generating extensive documentation, such as technical specifications or comprehensive reports that frequently exceed 1,000 lines. By breaking content into manageable sections with unique markers and integrity hashes, it prevents the data corruption and silent failures typically caused by LLM context compaction or session timeouts. This skill ensures continuous progress through robust initialization, status tracking, and automated repair commands, allowing users to resume generation exactly where it left off without losing context or wasting tokens.

主要功能

Incremental section-based generation with YAML-tracked progress
Integrity verification using content hashing and start/end markers
Built-in templates for B-SPEC and standard technical reports
Automatic resume capability for interrupted or context-limited sessions
1 GitHub stars
Automated cleanup that strips metadata for production-ready output

使用场景

Generating large-scale technical specifications (B-SPEC) exceeding 1,500 lines
Preventing content corruption when working near LLM context limits
Writing multi-section research papers or detailed analysis reports

关于

主要功能

Incremental section-based generation with YAML-tracked progress
Integrity verification using content hashing and start/end markers
Built-in templates for B-SPEC and standard technical reports
Automatic resume capability for interrupted or context-limited sessions
1 GitHub stars
Automated cleanup that strips metadata for production-ready output

使用场景

Generating large-scale technical specifications (B-SPEC) exceeding 1,500 lines
Preventing content corruption when working near LLM context limits
Writing multi-section research papers or detailed analysis reports