What is context-degradation?
context-degradation is a Claude skill by muratcankoylan that recognize patterns of context failure: lost-in-middle, poisoning, distraction, and clash. It is listed under Marketing & growth and has 18,159 GitHub stars.
| Type | Claude skill |
|---|---|
| Author | muratcankoylan |
| Repository | muratcankoylan/Agent-Skills-for-Context-Engineering |
| Popularity | 18,159 GitHub stars |
| Category | Marketing & growth |
| Security | Low risk (100/100) |
| Rank | #323 of 1,262 Claude skills; #28 in Marketing & growth |
| SKILL.md length | 236 lines |
| License | MIT |
| Last commit | 2026-10-12 |
| Stars this week | +52 |
Is context-degradation safe to install?
context-degradation scores 100/100 on the MCPMods security check: Low risk. No checks raised a concern. It passed 9 of 9 checks.
Skill results come from a static scan of its SKILL.md for instruction overrides, hidden actions, risky commands and credential access.
| Check | Result |
|---|---|
| No instruction-override phrases | Pass |
| No hidden actions | Pass |
| No pipe-to-shell installers | Pass |
| No obfuscated code | Pass |
| No destructive commands | Pass |
| Doesn't ask for keys or credentials | Pass |
| No network requests | Pass |
| No shell commands | Pass |
| Installs no packages | Pass |
How do I install context-degradation?
- Install with the skills CLI. Works for Claude Code and other agents that read skills.
npx skills add https://github.com/muratcankoylan/Agent-Skills-for-Context-Engineering --skill context-degradation
- Or copy it by hand. Claude Code loads any folder with a SKILL.md from ~/.claude/skills.
git clone https://github.com/muratcankoylan/Agent-Skills-for-Context-Engineering cp -r Agent-Skills-for-Context-Engineering/skills/context-degradation ~/.claude/skills/context-degradation
- Use it. Claude loads the skill when a task matches its description. You can also name it in your prompt.
Use the context-degradation skill to …
Or ask Claude to set it up
Paste this into Claude Code and it installs context-degradation for you:
Install the "context-degradation" skill from https://github.com/muratcankoylan/Agent-Skills-for-Context-Engineering/tree/main/skills/context-degradation into ~/.claude/skills, then tell me in one line what it does.
More from muratcankoylan/Agent-Skills-for-Context-Engineering
This repository has 8 Claude skills on MCPMods.
- context-fundamentals: Understand what context is, why it matters, and the anatomy of context in agent systems (Low risk)
- context-compression: Design and evaluate compression strategies for long-running sessions (Low risk)
- context-optimization: Apply compaction, masking, and caching strategies (Low risk)
- multi-agent-patterns: Master orchestrator, peer-to-peer, and hierarchical multi-agent architectures (Low risk)
- memory-systems: Design short-term, long-term, and graph-based memory architectures (Low risk)
- tool-design: Build tools that agents can use effectively, including architectural reduction patterns (Low risk)
- evaluation: Build evaluation frameworks for agent systems (Low risk)