guide
Protocol for the ctx-guard context and usage guard, and how to configure it. Use whenever a message starting with "[ctx-guard]" arrives (a context or usage note, a prompt turn saying the user is away, a refused ready token, a blocked tool, a lifted brake, a subagent budget note), when asked to write a ctx-guard handoff or end a reply with the ready token (CTX-READY), and when the user asks to set up, explain or change ctx-guard, a context guard or usage guard, its roles, thresholds, triggers, handoff, auto clear or compact, subagent budgets, trust or the /ctx-guard command.
Target: Target not categorized. Automatically collected · unverified.
Excerpt from the author’s documentation
Extracted from a fixed version of SKILL.md in its original language. This is the author’s documentation; installation, compatibility and runtime results have not been verified by this site.
Read the original English text
Topics in the source
- ctx-guard
- When a [ctx-guard] note arrives
- When a prompt says the user is away
- When the token is refused
- When a tool is blocked
- When a brake lifts
- When you are a subagent and get a budget note
- After /clear
ctx-guard watches the session's context fill and usage limits. It sends [ctx-guard] notes, blocks tools near limits, and, at a ready token, runs a guarded /clear plus resume (or /compact). This skill tells you how to answer its notes and how to help the operator configure it. Every key and default is in references/config-reference.md. The note states the figure, what to do now (by level), and how to end. Follow it in this order. 1. Finish the current step, as the note's level says: soft: finish the item in hand and keep new work small (for usage: fewer subagents, no speculative work); hard: start no new item, dispatch only what the current item needs, and bring it to a safe point; stop: dispatch nothing new and start no new work. 2. Reach a safe point. No subagents or background tasks running: wait for them or collect their results. Current work recorded: commit it, or write it down in the handoff. 3. Write the handoff, if the note asks for one. The note names the file (write the handoff to ) and quotes the role's instructions; follow them. The instructions name a skill: use it, and make sure the result ends up at that path. They name none: rewrite the whole file with the Write too
This is an excerpt. Refer to the complete source and its current requirements.
- Source repository
- svd/agent-skills
- Source file
- plugins/ctx-guard/skills/guide/SKILL.md
- Fixed revision
- cb47e43decd9132b8b9ffc29a43960cc5724435e
Requirements
- Only the author’s Skill metadata is included. Read the full instructions to check dependencies, licensing and conditions of use.
A small first task
先阅读原作者说明,再决定是否在自己的 AI 工具中使用。
Instructions for your AI tool
- Claude Desktop: resource-specific guide not available yet
- Claude Code: resource-specific guide not available yet
- Cursor: resource-specific guide not available yet
- Codex: resource-specific guide not available yet
Author documentation is not live verification by this site. Without evidence, we do not mark a resource as compatible, safe or free.
Sources and evidence
This record was collected automatically. Its documentation, suitability and runtime behavior have not been manually verified.
Author: svd/agent-skills · original repository metadata. Documentation review: Not reviewed. Last source check: 2026-11-11T00:00:19.058Z.
Categories are inferred from names, descriptions and paths. They do not establish compatibility.
License and attribution
MIT. Third-party content and copyright remain with their authors. Refer to the original license and terms.
Original license ↗Synchronization, copying and user-reported success do not mark a resource as runtime-tested.
Limitations and notes
- Collecting metadata does not mean this site has installed, adapted, runtime-tested or security-reviewed the resource.