claude-expert
Provides expert guidance on Claude prompting, skill creation, and memory management for enhanced AI agent performance.
Install this skill
or
85/100
Security score
The claude-expert skill was audited on Apr 14, 2026 and we found 3 security issues across 1 threat category. Review the findings below before installing.
Categories Tested
Security Issues
medium line 58
Access to home directory dotfiles
SourceSKILL.md
| 56 | |
| 57 | ### MCP |
| 58 | - Servers defined in `~/.claude.json`, `.mcp.json`, or managed-mcp.json |
| 59 | - Provide external tools and resources |
| 60 | - Use stdio, SSE, or HTTP transport |
medium line 63
Access to home directory dotfiles
SourceSKILL.md
| 61 | |
| 62 | ### Memory |
| 63 | - **Auto memory**: Claude automatically saves learnings to `~/.claude/projects/<project>/memory/` |
| 64 | - **MEMORY.md**: First 200 lines loaded at startup; topic files loaded on-demand |
| 65 | - **CLAUDE.md hierarchy**: managed policy > project > user > local (more specific wins) |
medium line 74
Access to home directory dotfiles
SourceSKILL.md
| 72 | - **Agent teams** (experimental): multiple independent sessions, shared task list, direct messaging |
| 73 | - Default to single agents; use multi-agent only for context protection, parallelization, or specialization |
| 74 | - Custom agents in `.claude/agents/` or `~/.claude/agents/` |
| 75 | |
| 76 | ### Workflows |
Scanned on Apr 14, 2026
View Security Dashboard