Documentation
¶
Overview ¶
Package mcpserver exposes gocrawl over the Model Context Protocol so agentic tools such as Conductor or Claude Code can drive crawls directly. It registers two tools: "crawl" (run a crawl + analysis and return a structured report) and "list_analyzers".
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
Types ¶
type CrawlInput ¶
type CrawlInput = crawlrequest.Params
CrawlInput is the MCP "crawl" tool input. Optional fields override the defaults. It is an alias for crawlrequest.Params, the mapping shared with the web API.
type CrawlOutput ¶
type CrawlOutput struct {
Report CrawlReport `json:"report"`
}
CrawlOutput is the MCP "crawl" tool output.
type CrawlReport ¶ added in v0.7.0
type CrawlReport struct {
Seed string `json:"seed"`
StartedAt string `json:"started_at"`
FinishedAt string `json:"finished_at"`
PagesCrawled int `json:"pages_crawled"`
Summary report.Summary `json:"summary"`
Issues []analyze.Issue `json:"issues"`
Notes []string `json:"notes,omitempty"`
Coverage *crawler.Coverage `json:"coverage,omitempty"`
}
CrawlReport mirrors report.Report for the MCP "crawl" tool, minus SiteMap. SiteMap's *sitemapgen.Node is self-referential (Children []*Node), and the go-sdk schema reflector panics on cyclic types when building the tool's output schema. Agents reason over the flat Issues list, not the tree, so it's dropped here; the CLI/web JSON report still includes it.
type ListAnalyzersInput ¶
type ListAnalyzersInput struct{}
ListAnalyzersInput is the (empty) input for the "list_analyzers" tool.
type ListAnalyzersOutput ¶
type ListAnalyzersOutput struct {
Analyzers []runner.AnalyzerInfo `json:"analyzers"`
}
ListAnalyzersOutput lists the available analyzers.
Source Files