dbengoa42-star/edgeclaw/compact-guide

Claude Code and agent skill by dbengoa42-star/edgeclaw

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Behavioral guidance for long conversations with context compaction. Follow these rules to avoid redundant work after context is compressed.

Install

npx vskill@latest install dbengoa42-star/edgeclaw/compact-guide

SKILL.md

The same SKILL.md was published earlier as dbengoa42-star/edgeclaw/compact-guide.

Long Conversation Behavior Guide

After context compaction, some older messages are replaced by a summary. Follow these rules to maintain quality:

Rules

  1. Trust the summary. If a [Previous Context Summary] section is present in your system prompt, it accurately describes what happened before. Do not ask the user to repeat information that is in the summary.
  1. Do not repeat completed work. If the summary says a task was completed, do not redo it. If you need to verify, use a tool to check the result (e.g., read the file, check the state) rather than redoing the entire operation.
  1. Recent tool summaries are reliable. The [Recent Tool Result Summaries] section contains compressed versions of recent long tool outputs. Reference these instead of re-running the same tool calls.
  1. When in doubt, inspect. Use clawxgovernor__context_inspect to check what the context engine currently sees. This is better than guessing.
  1. Warn about approaching limits. If you notice estimatedRemainingTurns is low (< 3), proactively tell the user that context is getting full and suggest wrapping up or starting a fresh session.
  1. Do not fight compaction. If old messages are gone, they are gone intentionally. Work with the summary rather than trying to reconstruct the original conversation.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/dbengoa42-star/edgeclaw/compact-guide/badge)](https://verified-skill.com/skills/dbengoa42-star/edgeclaw/compact-guide)

Something wrong with this skill or its page? Tell us.