skills.rest
SukbeomH
SukbeomHCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

context-health-monitor

Monitor agent context health and dump session state to .hxsk/STATE.md.

2|Updated Jan 12, 2026
Search Tags:#debugging#context management#state management#agent health#session quality#proactive monitoring
One-click install
npx skills add https://github.com/SukbeomH/HExoskeleton --skill context-health-monitor-sukbeomh
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: context-health-monitor
Source: https://github.com/SukbeomH/HExoskeleton/tree/main/.claude/skills/context-health-monitor
Command: npx skills add https://github.com/SukbeomH/HExoskeleton --skill context-health-monitor-sukbeomh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill prevents "Context Rot" by proactively monitoring the agent's context window usage and session quality, triggering state dumps and recommending actions before performance degrades.

Core Features & Use Cases

  • Proactive Compaction: Runs /compact before the context window is full to preserve information.
  • 3-Strike Rule: Stops attempts and dumps state after 3 failed debugging attempts on the same issue.
  • Circular Detection: Identifies repetitive approaches and suggests alternatives or pauses.
  • Use Case: If an agent is stuck trying to fix a bug and fails three times, this skill will save the current state, document what was tried, and suggest starting a fresh session to avoid further degradation.

Quick Start

Use the context-health-monitor skill to dump the current session state to .hxsk/STATE.md.

Frequently Asked Questions about context-health-monitor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent context rot during long agent debugging sessions?▼

Prevent context rot by proactively monitoring agent context window usage and session quality, triggering state dumps and recommending actions before performance degrades. It tracks recurring issues and compacts data to preserve information.

What is the 3-strike rule for debugging failures in agent state management?▼

The 3-strike rule stops debugging attempts and dumps state after 3 failed tries on the same issue. It documents what was tried and suggests starting a fresh session to avoid further context degradation.

How do I detect circular approaches when an agent is stuck on a bug?▼

Detect circular approaches by monitoring agent session quality for repetitive debugging patterns. The skill identifies when an agent is stuck and suggests alternative approaches or pauses to prevent context rot.

How do I save agent session state to prevent data loss?▼

Save agent session state by dumping the current context to memory files like .hxsk/STATE.md. This stores session state and recurring issues for later analysis and recovery.

When should I run proactive compaction on an agent context window?▼

Run proactive compaction before the context window is full to preserve information. The skill monitors usage and triggers compaction automatically to maintain session quality and prevent performance degradation.

Related Skills

Explore More Software Engineering Skills→
antoinerswantoinersw

setup-reactive

Configure the Clean Claude Reactive System with state storage, routing hooks, and agent communication scripts.

Community
Intermediate
cameronfreercameronfreer
360

lean4-memories

Persist Lean 4 proof patterns and failed approaches across sessions via MCP memory server.

Community
Advanced
Rune-kitRune-kit
82

context-engine

Monitor tool call counts and trigger state saving before context compaction.

Official
Advanced
CobaltSatoCobaltSato

strategic-compact

Prompt manual /compact at logical workflow boundaries using configurable thresholds.

Community
Advanced