tao-context-vcc
CommunityShrink TAO transcripts with zero token cost.
System Documentation
What problem does it solve?
This Skill prevents TAO sessions from running out of context by deterministically compacting long, repetitive transcripts before they reach the model.
Core Features & Use Cases
- Tool-output dedup: Collapses repeated tool turns into a truncated placeholder to remove identical noise.
- Verbose-block head/tail truncation: Clips extremely large blocks by keeping the beginning and end while marking omitted content.
- Repeat-pattern collapse: Detects runs of identical messages and replaces them with a repetition counter.
- Whitespace + log-noise normalization: Normalizes CRLF to LF, removes trailing whitespace, and deduplicates blank lines to reduce churn.
Real-world use: When a TAO loop repeatedly streams similar tool outputs during repository analysis, compacting the transcript can keep the session within the active model context window.
Quick Start
Run compact on your saved TAO messages to produce a smaller, idempotent transcript: call compact(messages, target_token_budget=120_000) and forward the returned compacted messages to your next TAO step.
Dependency Matrix
Required Modules
None requiredComponents
Standard package💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: tao-context-vcc Download link: https://github.com/CleanExpo/Pi-Dev-Ops/archive/main.zip#tao-context-vcc Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.