Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
MCP Config
v0.1.0MCP server configuration (.mcp.json) management. add - add server (scope selection + registration) [add.md], format - JSON format reference (differences per...
⭐ 0· 34·0 current·0 all-time
byes6kr@drumrobot
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (MCP server config management) match the instructions and files: all guidance is about adding/formatting/cataloging .mcp.json entries and selecting global vs project scope.
Instruction Scope
SKILL.md and add.md explicitly tell the agent to read/modify project and global config files (./.mcp.json, ~/.mcp.json, ~/.cursor/mcp.json, ~/.gemini/antigravity/mcp_config.json) and to prompt the user for scope/agent selection. That is coherent with a config-manager, but it means the skill will direct writes to other agents' config files and can cause cross-agent effects — the user should expect and review any global changes.
Install Mechanism
No install spec and no code files — instruction-only — so nothing is downloaded or written by an installer as part of skill activation.
Credentials
The skill declares no required env vars or credentials (consistent). However, examples show storing secrets (DATABASE_URI, API_KEY) and using interpolations like ${API_KEY}. This is expected for a config tool but increases risk if users paste credentials into global configs or allow untrusted entries.
Persistence & Privilege
always is false and there is no install. The only persisted effect is editing config files (project or per-agent global files) which is consistent with the stated purpose.
Assessment
This skill is coherent for managing .mcp.json but exercise caution before applying changes, especially to global config files: 1) Prefer project-scoped additions unless you trust the change for all agents. 2) Review any package names/commands (e.g., npx @package, uvx) before adding — these entries will cause code to run when the MCP is invoked. 3) Avoid placing plaintext credentials in global configs; use secure storage or environment injection where possible. 4) Pay attention to examples like uvx --access-mode=unrestricted and DATABASE_URI with credentials — they grant broad access and should only be used for trusted services. 5) Confirm the user prompts (AskUserQuestion) are actually used to get explicit consent before writing global files.Like a lobster shell, security has layers — review code before you run it.
latestvk97dmfnd87n9gag0xpxxqx4df9841a0r
License
MIT-0
Free to use, modify, and redistribute. No attribution required.