LogoClawIndex
CasesSkillsAbout
LogoClawIndex

self-managed-context - Manage a model-controlled editable context window

Enables a model to retain, compact, evict, and rewrite live context through an editable file while the harness enforces safety, budgeting, and rollback.

Tags

Updated: 2026-10-19

Capabilities

Typical Inputs

Typical Outputs

What this skill does

  • Manage editable live context
  • Render context to a file
  • Edit context with code tools
  • Parse files into messages
  • Pin system and task messages
  • Apply fit and shrink gates
  • Record edit receipts
  • Show token budget readouts
  • Roll back turns on overflow
  • Steer context-editing strategy

Inputs

  • Live conversation context
  • System prompt
  • Task messages
  • Shell and code tools
  • Token budget data
  • Tokenizer configuration

Outputs

  • Editable context file
  • Parsed message list
  • Edit receipts
  • Token budget readouts
  • Rollback status

Requirements

  • Model tool-use support
  • Read-write access to live context
  • Harness-controlled pinned prefix
  • Tolerant message parsing
  • Local token counting
  • Overflow rollback support

Source

  • Spec: SKILL.md

ClawIndex

OpenClaw Skills & Use Case Index

ClawIndex is an ecosystem-driven index of OpenClaw skills and real-world use cases.

Index

Skills·
Cases

Meta

About·
Disclaimer·
Email·
GitHub
© 2026 ClawIndex All Rights Reserved.
context management
editable context
agent harness
prefix caching
context compaction
model-controlled agents
Manage editable live context
Render context to a file
Edit context with code tools
Parse files into messages
Live conversation context
System prompt
Task messages
Editable context file
Parsed message list
Edit receipts