Streams long-form documents to markdown files with incremental saving and automatic recovery capabilities.
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.
Key Features
01Incremental section-based generation with YAML-tracked progress
02Integrity verification using content hashing and start/end markers
03Built-in templates for B-SPEC and standard technical reports
04Automatic resume capability for interrupted or context-limited sessions
051 GitHub stars
06Automated cleanup that strips metadata for production-ready output
