MCP Filesystem Server

search_files

Find files and directories by name pattern across subdirectories. Search recursively from a starting path with case-insensitive matching for partial names.

Instructions

Recursively search for files and directories matching a pattern. Searches through all subdirectories from the starting path. The search is case-insensitive and matches partial names. Returns full paths to all matching items. Great for finding files when you don't know their exact location. Only searches within allowed directories.

Input Schema

TableJSON Schema

Name	Required	Description
`path`	Yes	Root path to start searching from
`pattern`	Yes	Pattern to match against filenames and directories
`excludePatterns`	No	Patterns to exclude from search results

Implementation Reference

src/index.ts:606-680 (handler)

The main handler for the 'search_files' tool. Validates input using SearchFilesArgsSchema, performs recursive search starting from the given path, matches filenames case-insensitively against the pattern, supports excludePatterns using minimatch globs, collects full paths of matches, and returns formatted results or 'No matches found'.

case 'search_files': {
  const parsed = SearchFilesArgsSchema.safeParse(a)
  if (!parsed.success) {
    throw new FileSystemError(`Invalid arguments for ${name}`, 'INVALID_ARGS', undefined, {
      errors: parsed.error.format(),
    })
  }

  const validPath = await validatePath(parsed.data.path, config)
  const patternLower = parsed.data.pattern.toLowerCase()
  const results: string[] = []

  async function search(currentPath: string) {
    try {
      const entries = await fs.readdir(currentPath, { withFileTypes: true })

      for (const entry of entries) {
        const fullPath = path.join(currentPath, entry.name)

        try {
          await validatePath(fullPath, config)
          const relativePath = path.relative(validPath, fullPath)

          // Check if the path should be excluded
          const shouldExclude =
            parsed.data &&
            parsed.data.excludePatterns.some((excludePattern) => {
              const globPattern = excludePattern.includes('*')
                ? excludePattern
                : `**/${excludePattern}**`
              return minimatch(relativePath, globPattern, { nocase: true })
            })

          if (shouldExclude) {
            continue
          }

          // Check if the name matches the search pattern
          if (entry.name.toLowerCase().includes(patternLower)) {
            results.push(fullPath)
          }

          // Recursively search subdirectories
          if (entry.isDirectory()) {
            await search(fullPath)
          }
        } catch (error) {
          // Skip paths we can't access or validate
          continue
        }
      }
    } catch (error) {
      // Skip directories we can't read
      return
    }
  }

  await search(validPath)
  await logger.debug(`Search complete: ${parsed.data.pattern}`, {
    resultCount: results.length,
  })

  endMetric()
  return {
    content: [
      {
        type: 'text',
        text:
          results.length > 0
            ? `Found ${results.length} matches:\n${results.join('\n')}`
            : 'No matches found',
      },
    ],
  }
}

src/index.ts:133-141 (schema)

Zod schema defining the input parameters for the search_files tool: required path (root to search from), pattern (string to match in names), and optional excludePatterns array.

const SearchFilesArgsSchema = z.object({
  path: z.string().describe('Root path to start searching from'),
  pattern: z.string().describe('Pattern to match against filenames and directories'),
  excludePatterns: z
    .array(z.string())
    .optional()
    .default([])
    .describe('Patterns to exclude from search results'),
})

src/index.ts:306-315 (registration)

Registers the search_files tool in the ListTools response, providing name, detailed description, and inputSchema converted from SearchFilesArgsSchema.

{
  name: 'search_files',
  description:
    'Recursively search for files and directories matching a pattern. ' +
    'Searches through all subdirectories from the starting path. The search ' +
    'is case-insensitive and matches partial names. Returns full paths to all ' +
    "matching items. Great for finding files when you don't know their exact location. " +
    'Only searches within allowed directories.',
  inputSchema: zodToJsonSchema(SearchFilesArgsSchema) as ToolInput,
},

Schema Changelog

Changes observed during successful MCP inspections.

First observedAug 22, 2025

TDQS

A4/5.0

Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It effectively describes key behaviors: recursive search, case-insensitivity, partial name matching, and restriction to allowed directories. However, it does not cover aspects like performance implications (e.g., time for deep searches), error handling (e.g., invalid paths), or output format details (e.g., how results are structured), leaving some gaps for a tool with no annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is front-loaded with the core functionality in the first sentence, followed by supporting details in concise sentences. Each sentence adds value: recursive nature, search scope, matching behavior, return value, usage scenario, and constraints. There is no wasted text, making it highly efficient and well-structured.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's moderate complexity (3 parameters, no output schema, no annotations), the description is largely complete. It covers purpose, behavior, and constraints adequately. However, without an output schema, it could benefit from more detail on the return format (e.g., structure of results, error cases), slightly reducing completeness for a search tool with no structured output documentation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema description coverage is 100%, so the schema already documents all parameters (path, pattern, excludePatterns) with clear descriptions. The description adds no additional parameter-specific information beyond what the schema provides, such as examples of pattern syntax or details on excludePatterns usage. Thus, it meets the baseline of 3 without compensating further.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose with specific verbs ('search for files and directories matching a pattern') and resources ('files and directories'), distinguishing it from siblings like list_directory (which lists without searching) or get_file_info (which gets metadata for a known file). It explicitly mentions the recursive nature and case-insensitive partial matching, which further differentiates its functionality.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides clear context for when to use this tool ('Great for finding files when you don't know their exact location') and mentions constraints ('Only searches within allowed directories'), but it does not explicitly name alternatives or specify when not to use it compared to siblings like list_directory or directory_tree. The guidance is helpful but lacks explicit sibling differentiation.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server