skills_search
AI agents call skills_search to retrieve information from Skills MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool name 'search' indicates a query or discovery operation with no side effects. It retrieves information about available skills from the registry without modifying, executing, or deleting anything. The empty description lowers confidence slightly, but the naming convention and sibling tools (list, get_details) confirm this is a Read operation typical of package manager search functionality.
From the tool's definition Tool name is 'skills_search' with an empty description. Based on the server context as a 'package manager for AI agents' with sibling tools including 'skills_list' and 'skills_get_details', this follows the standard pattern of discovery/query operations in…
Attacks that exploit this kind of access
skills_search. It is categorised as a Read tool in the Skills MCP MCP Server, which means it retrieves data without modifying state.
Register the Skills MCP server in PolicyLayer and add a rule for skills_search: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Skills MCP. Nothing to install.
skills_search is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the skills_search rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for skills_search. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
skills_search is provided by the Skills MCP server (leezhuuuuu/skills-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →