bonecode 1.0.0

package/src/engine/tool/tool/webfetch.ts

@@ -0,0 +1,192 @@
+import { Effect, Schema } from "effect"
+import { HttpClient, HttpClientRequest } from "effect/unstable/http"
+import { Parser } from "htmlparser2"
+import * as Tool from "./tool"
+import TurndownService from "turndown"
+import DESCRIPTION from "./webfetch.txt"
+import { isImageAttachment } from "@/util/media"
+
+const MAX_RESPONSE_SIZE = 5 * 1024 * 1024 // 5MB
+const DEFAULT_TIMEOUT = 30 * 1000 // 30 seconds
+const MAX_TIMEOUT = 120 * 1000 // 2 minutes
+
+export const Parameters = Schema.Struct({
+  url: Schema.String.annotate({ description: "The URL to fetch content from" }),
+  format: Schema.Literals(["text", "markdown", "html"])
+    .annotate({
+      description: "The format to return the content in (text, markdown, or html). Defaults to markdown.",
+      default: "markdown",
+    })
+    .pipe(Schema.optional, Schema.withDecodingDefault(Effect.succeed("markdown" as const))),
+  timeout: Schema.optional(Schema.Number).annotate({ description: "Optional timeout in seconds (max 120)" }),
+})
+
+export const WebFetchTool = Tool.define(
+  "webfetch",
+  Effect.gen(function* () {
+    const http = yield* HttpClient.HttpClient
+    const httpOk = HttpClient.filterStatusOk(http)
+
+    return {
+      description: DESCRIPTION,
+      parameters: Parameters,
+      execute: (params: Schema.Schema.Type<typeof Parameters>, ctx: Tool.Context) =>
+        Effect.gen(function* () {
+          if (!params.url.startsWith("http://") && !params.url.startsWith("https://")) {
+            throw new Error("URL must start with http:// or https://")
+          }
+
+          yield* ctx.ask({
+            permission: "webfetch",
+            patterns: [params.url],
+            always: ["*"],
+            metadata: {
+              url: params.url,
+              format: params.format,
+              timeout: params.timeout,
+            },
+          })
+
+          const timeout = Math.min((params.timeout ?? DEFAULT_TIMEOUT / 1000) * 1000, MAX_TIMEOUT)
+
+          // Build Accept header based on requested format with q parameters for fallbacks
+          let acceptHeader = "*/*"
+          switch (params.format) {
+            case "markdown":
+              acceptHeader = "text/markdown;q=1.0, text/x-markdown;q=0.9, text/plain;q=0.8, text/html;q=0.7, */*;q=0.1"
+              break
+            case "text":
+              acceptHeader = "text/plain;q=1.0, text/markdown;q=0.9, text/html;q=0.8, */*;q=0.1"
+              break
+            case "html":
+              acceptHeader =
+                "text/html;q=1.0, application/xhtml+xml;q=0.9, text/plain;q=0.8, text/markdown;q=0.7, */*;q=0.1"
+              break
+            default:
+              acceptHeader =
+                "text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8"
+          }
+          const headers = {
+            "User-Agent":
+              "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/143.0.0.0 Safari/537.36",
+            Accept: acceptHeader,
+            "Accept-Language": "en-US,en;q=0.9",
+          }
+
+          const request = HttpClientRequest.get(params.url).pipe(HttpClientRequest.setHeaders(headers))
+
+          // Retry with honest UA if blocked by Cloudflare bot detection (TLS fingerprint mismatch)
+          const response = yield* httpOk.execute(request).pipe(
+            Effect.catchIf(
+              (err) =>
+                err.reason._tag === "StatusCodeError" &&
+                err.reason.response.status === 403 &&
+                err.reason.response.headers["cf-mitigated"] === "challenge",
+              () =>
+                httpOk.execute(
+                  HttpClientRequest.get(params.url).pipe(
+                    HttpClientRequest.setHeaders({ ...headers, "User-Agent": "opencode" }),
+                  ),
+                ),
+            ),
+            Effect.timeoutOrElse({ duration: timeout, orElse: () => Effect.die(new Error("Request timed out")) }),
+          )
+
+          // Check content length
+          const contentLength = response.headers["content-length"]
+          if (contentLength && parseInt(contentLength) > MAX_RESPONSE_SIZE) {
+            throw new Error("Response too large (exceeds 5MB limit)")
+          }
+
+          const arrayBuffer = yield* response.arrayBuffer
+          if (arrayBuffer.byteLength > MAX_RESPONSE_SIZE) {
+            throw new Error("Response too large (exceeds 5MB limit)")
+          }
+
+          const contentType = response.headers["content-type"] || ""
+          const mime = contentType.split(";")[0]?.trim().toLowerCase() || ""
+          const title = `${params.url} (${contentType})`
+
+          if (isImageAttachment(mime)) {
+            const base64Content = Buffer.from(arrayBuffer).toString("base64")
+            return {
+              title,
+              output: "Image fetched successfully",
+              metadata: {},
+              attachments: [
+                {
+                  type: "file" as const,
+                  mime,
+                  url: `data:${mime};base64,${base64Content}`,
+                },
+              ],
+            }
+          }
+
+          const content = new TextDecoder().decode(arrayBuffer)
+
+          // Handle content based on requested format and actual content type
+          switch (params.format) {
+            case "markdown":
+              if (contentType.includes("text/html")) {
+                const markdown = convertHTMLToMarkdown(content)
+                return {
+                  output: markdown,
+                  title,
+                  metadata: {},
+                }
+              }
+              return { output: content, title, metadata: {} }
+
+            case "text":
+              if (contentType.includes("text/html")) {
+                return { output: extractTextFromHTML(content), title, metadata: {} }
+              }
+              return { output: content, title, metadata: {} }
+
+            case "html":
+              return { output: content, title, metadata: {} }
+
+            default:
+              return { output: content, title, metadata: {} }
+          }
+        }).pipe(Effect.orDie),
+    }
+  }),
+)
+
+function extractTextFromHTML(html: string) {
+  let text = ""
+  let skipDepth = 0
+
+  const parser = new Parser({
+    onopentag(name) {
+      if (skipDepth > 0 || ["script", "style", "noscript", "iframe", "object", "embed"].includes(name)) {
+        skipDepth++
+      }
+    },
+    ontext(input) {
+      if (skipDepth === 0) text += input
+    },
+    onclosetag() {
+      if (skipDepth > 0) skipDepth--
+    },
+  })
+
+  parser.write(html)
+  parser.end()
+
+  return text.trim()
+}
+
+function convertHTMLToMarkdown(html: string): string {
+  const turndownService = new TurndownService({
+    headingStyle: "atx",
+    hr: "---",
+    bulletListMarker: "-",
+    codeBlockStyle: "fenced",
+    emDelimiter: "*",
+  })
+  turndownService.remove(["script", "style", "meta", "link"])
+  return turndownService.turndown(html)
+}

package/src/engine/tool/tool/webfetch.txt

@@ -0,0 +1,13 @@
+- Fetches content from a specified URL
+- Takes a URL and optional format as input
+- Fetches the URL content, converts to requested format (markdown by default)
+- Returns the content in the specified format
+- Use this tool when you need to retrieve and analyze web content
+
+Usage notes:
+  - IMPORTANT: if another tool is present that offers better web fetching capabilities, is more targeted to the task, or has fewer restrictions, prefer using that tool instead of this one.
+  - The URL must be a fully-formed valid URL
+  - HTTP URLs will be automatically upgraded to HTTPS
+  - Format options: "markdown" (default), "text", or "html"
+  - This tool is read-only and does not modify any files
+  - Results may be summarized if the content is very large

package/src/engine/tool/tool/websearch.ts

@@ -0,0 +1,143 @@
+import { Effect, Schema } from "effect"
+import { HttpClient } from "effect/unstable/http"
+import * as Tool from "./tool"
+import * as McpWebSearch from "./mcp-websearch"
+import DESCRIPTION from "./websearch.txt"
+import { checksum } from "@opencode-ai/core/util/encode"
+import { InstallationVersion } from "@opencode-ai/core/installation/version"
+import { RuntimeFlags } from "@/effect/runtime-flags"
+
+export const Parameters = Schema.Struct({
+  query: Schema.String.annotate({ description: "Websearch query" }),
+  numResults: Schema.optional(Schema.Number).annotate({
+    description: "Number of search results to return (default: 8)",
+  }),
+  livecrawl: Schema.optional(Schema.Literals(["fallback", "preferred"])).annotate({
+    description:
+      "Live crawl mode - 'fallback': use live crawling as backup if cached content unavailable, 'preferred': prioritize live crawling (default: 'fallback')",
+  }),
+  type: Schema.optional(Schema.Literals(["auto", "fast", "deep"])).annotate({
+    description: "Search type - 'auto': balanced search (default), 'fast': quick results, 'deep': comprehensive search",
+  }),
+  contextMaxCharacters: Schema.optional(Schema.Number).annotate({
+    description: "Maximum characters for context string optimized for LLMs (default: 10000)",
+  }),
+})
+
+const WebSearchProviderSchema = Schema.Literals(["exa", "parallel"])
+export type WebSearchProvider = Schema.Schema.Type<typeof WebSearchProviderSchema>
+
+export function selectWebSearchProvider(sessionID: string, flags = { exa: false, parallel: false }): WebSearchProvider {
+  const override = process.env.OPENCODE_WEBSEARCH_PROVIDER
+  if (override === "exa" || override === "parallel") return override
+  if (flags.parallel) return "parallel"
+  if (flags.exa) return "exa"
+
+  return Number.parseInt(checksum(sessionID) ?? "0", 36) % 2 === 0 ? "exa" : "parallel"
+}
+
+export function webSearchProviderLabel(provider: unknown) {
+  if (provider === "parallel") return "Parallel Web Search"
+  if (provider === "exa") return "Exa Web Search"
+  return "Web Search"
+}
+
+export function webSearchModelName(extra: Tool.Context["extra"]) {
+  const model = extra?.model
+  if (!model || typeof model !== "object") return undefined
+  const api = "api" in model && model.api && typeof model.api === "object" ? model.api : undefined
+  const apiID = api && "id" in api && typeof api.id === "string" ? api.id : undefined
+  const id = "id" in model && typeof model.id === "string" ? model.id : undefined
+  return (apiID ?? id)?.slice(0, 100)
+}
+
+function parallelAuthHeaders() {
+  const headers = { "User-Agent": `opencode/${InstallationVersion}` }
+  if (!process.env.PARALLEL_API_KEY) return headers
+  return { ...headers, Authorization: `Bearer ${process.env.PARALLEL_API_KEY}` }
+}
+
+function callProvider(
+  http: HttpClient.HttpClient,
+  provider: WebSearchProvider,
+  params: Schema.Schema.Type<typeof Parameters>,
+  ctx: Tool.Context,
+) {
+  if (provider === "parallel") {
+    return McpWebSearch.call(
+      http,
+      McpWebSearch.PARALLEL_URL,
+      "web_search",
+      McpWebSearch.ParallelSearchArgs,
+      {
+        objective: params.query,
+        search_queries: [params.query],
+        session_id: ctx.sessionID,
+        model_name: webSearchModelName(ctx.extra),
+      },
+      "25 seconds",
+      parallelAuthHeaders(),
+    )
+  }
+
+  return McpWebSearch.call(
+    http,
+    McpWebSearch.EXA_URL,
+    "web_search_exa",
+    McpWebSearch.SearchArgs,
+    {
+      query: params.query,
+      type: params.type || "auto",
+      numResults: params.numResults || 8,
+      livecrawl: params.livecrawl || "fallback",
+      contextMaxCharacters: params.contextMaxCharacters,
+    },
+    "25 seconds",
+  )
+}
+
+export const WebSearchTool = Tool.define(
+  "websearch",
+  Effect.gen(function* () {
+    const http = yield* HttpClient.HttpClient
+    const flags = yield* RuntimeFlags.Service
+
+    return {
+      get description() {
+        return DESCRIPTION.replace("{{year}}", new Date().getFullYear().toString())
+      },
+      parameters: Parameters,
+      execute: (params: Schema.Schema.Type<typeof Parameters>, ctx: Tool.Context) =>
+        Effect.gen(function* () {
+          const provider = selectWebSearchProvider(ctx.sessionID, {
+            exa: flags.enableExa,
+            parallel: flags.enableParallel,
+          })
+          const title = webSearchProviderLabel(provider)
+          yield* ctx.metadata({ title: `${title} "${params.query}"`, metadata: { provider } })
+
+          yield* ctx.ask({
+            permission: "websearch",
+            patterns: [params.query],
+            always: ["*"],
+            metadata: {
+              query: params.query,
+              numResults: params.numResults,
+              livecrawl: params.livecrawl,
+              type: params.type,
+              contextMaxCharacters: params.contextMaxCharacters,
+              provider,
+            },
+          })
+
+          const result = yield* callProvider(http, provider, params, ctx)
+
+          return {
+            output: result ?? "No search results found. Please try a different query.",
+            title: `${title}: ${params.query}`,
+            metadata: { provider },
+          }
+        }).pipe(Effect.orDie),
+    }
+  }),
+)

package/src/engine/tool/tool/websearch.txt

@@ -0,0 +1,14 @@
+- Search the web using the session's web search provider - performs real-time web searches and can scrape content from specific URLs
+- Provides up-to-date information for current events and recent data
+- Supports configurable result counts and returns the content from the most relevant websites
+- Use this tool for accessing information beyond knowledge cutoff
+- Searches are performed automatically within a single API call
+
+Usage notes:
+  - Supports live crawling modes when available: 'fallback' (backup if cached unavailable) or 'preferred' (prioritize live crawling)
+  - Search types when available: 'auto' (balanced), 'fast' (quick results), 'deep' (comprehensive search)
+  - Configurable context length for optimal LLM integration
+  - Domain filtering and advanced search options available
+
+The current year is {{year}}. You MUST use this year when searching for recent information or current events
+- Example: If the current year is 2026 and the user asks for "latest AI news", search for "AI news 2026", NOT "AI news 2025"

package/src/engine/tool/tool/write.ts

@@ -0,0 +1,104 @@
+import { Schema } from "effect"
+import * as path from "path"
+import { Effect } from "effect"
+import * as Tool from "./tool"
+import { LSP } from "@/lsp/lsp"
+import { createTwoFilesPatch } from "diff"
+import DESCRIPTION from "./write.txt"
+import { Bus } from "../bus"
+import { File } from "../file"
+import { FileWatcher } from "../file/watcher"
+import { Format } from "../format"
+import { AppFileSystem } from "@opencode-ai/core/filesystem"
+import { InstanceState } from "@/effect/instance-state"
+import { trimDiff } from "./edit"
+import { assertExternalDirectoryEffect } from "./external-directory"
+import * as Bom from "@/util/bom"
+
+const MAX_PROJECT_DIAGNOSTICS_FILES = 5
+
+export const Parameters = Schema.Struct({
+  content: Schema.String.annotate({ description: "The content to write to the file" }),
+  filePath: Schema.String.annotate({
+    description: "The absolute path to the file to write (must be absolute, not relative)",
+  }),
+})
+
+export const WriteTool = Tool.define(
+  "write",
+  Effect.gen(function* () {
+    const lsp = yield* LSP.Service
+    const fs = yield* AppFileSystem.Service
+    const bus = yield* Bus.Service
+    const format = yield* Format.Service
+
+    return {
+      description: DESCRIPTION,
+      parameters: Parameters,
+      execute: (params: { content: string; filePath: string }, ctx: Tool.Context) =>
+        Effect.gen(function* () {
+          const instance = yield* InstanceState.context
+          const filepath = path.isAbsolute(params.filePath)
+            ? params.filePath
+            : path.join(instance.directory, params.filePath)
+          yield* assertExternalDirectoryEffect(ctx, filepath)
+
+          const exists = yield* fs.existsSafe(filepath)
+          const source = exists ? yield* Bom.readFile(fs, filepath) : { bom: false, text: "" }
+          const next = Bom.split(params.content)
+          const desiredBom = source.bom || next.bom
+          const contentOld = source.text
+          const contentNew = next.text
+
+          const diff = trimDiff(createTwoFilesPatch(filepath, filepath, contentOld, contentNew))
+          yield* ctx.ask({
+            permission: "edit",
+            patterns: [path.relative(instance.worktree, filepath)],
+            always: ["*"],
+            metadata: {
+              filepath,
+              diff,
+            },
+          })
+
+          yield* fs.writeWithDirs(filepath, Bom.join(contentNew, desiredBom))
+          if (yield* format.file(filepath)) {
+            yield* Bom.syncFile(fs, filepath, desiredBom)
+          }
+          yield* bus.publish(File.Event.Edited, { file: filepath })
+          yield* bus.publish(FileWatcher.Event.Updated, {
+            file: filepath,
+            event: exists ? "change" : "add",
+          })
+
+          let output = "Wrote file successfully."
+          yield* lsp.touchFile(filepath, "document")
+          const diagnostics = yield* lsp.diagnostics()
+          const normalizedFilepath = AppFileSystem.normalizePath(filepath)
+          let projectDiagnosticsCount = 0
+          for (const [file, issues] of Object.entries(diagnostics)) {
+            const current = file === normalizedFilepath
+            if (!current && projectDiagnosticsCount >= MAX_PROJECT_DIAGNOSTICS_FILES) continue
+            const block = LSP.Diagnostic.report(current ? filepath : file, issues)
+            if (!block) continue
+            if (current) {
+              output += `\n\nLSP errors detected in this file, please fix:\n${block}`
+              continue
+            }
+            projectDiagnosticsCount++
+            output += `\n\nLSP errors detected in other files:\n${block}`
+          }
+
+          return {
+            title: path.relative(instance.worktree, filepath),
+            metadata: {
+              diagnostics,
+              filepath,
+              exists: exists,
+            },
+            output,
+          }
+        }).pipe(Effect.orDie),
+    }
+  }),
+)

package/src/engine/tool/tool/write.txt

@@ -0,0 +1,8 @@
+Writes a file to the local filesystem.
+
+Usage:
+- This tool will overwrite the existing file if there is one at the provided path.
+- If this is an existing file, you MUST use the Read tool first to read the file's contents. This tool will fail if you did not read the file first.
+- ALWAYS prefer editing existing files in the codebase. NEVER write new files unless explicitly required.
+- NEVER proactively create documentation files (*.md) or README files. Only create documentation files if explicitly requested by the User.
+- Only use emojis if the user explicitly requests it. Avoid writing emojis to files unless asked.

package/src/engine/tool/tool.ts

@@ -0,0 +1,164 @@
+import { Effect, Schema } from "effect"
+import type { JSONSchema7 } from "@ai-sdk/provider"
+import type { MessageV2 } from "../session/message-v2"
+import type { Permission } from "../permission"
+import type { SessionID, MessageID } from "../session/schema"
+import * as Truncate from "./truncate"
+import { Agent } from "@/agent/agent"
+
+interface Metadata {
+  [key: string]: any
+}
+
+// TODO: remove this hack
+export type DynamicDescription = (agent: Agent.Info) => Effect.Effect<string>
+
+export type Context<M extends Metadata = Metadata> = {
+  sessionID: SessionID
+  messageID: MessageID
+  agent: string
+  abort: AbortSignal
+  callID?: string
+  extra?: { [key: string]: unknown }
+  messages: MessageV2.WithParts[]
+  metadata(input: { title?: string; metadata?: M }): Effect.Effect<void>
+  ask(input: Omit<Permission.Request, "id" | "sessionID" | "tool">): Effect.Effect<void>
+}
+
+export interface ExecuteResult<M extends Metadata = Metadata> {
+  title: string
+  metadata: M
+  output: string
+  attachments?: Omit<MessageV2.FilePart, "id" | "sessionID" | "messageID">[]
+}
+
+export interface Def<
+  Parameters extends Schema.Decoder<unknown> = Schema.Decoder<unknown>,
+  M extends Metadata = Metadata,
+> {
+  id: string
+  description: string
+  parameters: Parameters
+  jsonSchema?: JSONSchema7
+  execute(args: Schema.Schema.Type<Parameters>, ctx: Context): Effect.Effect<ExecuteResult<M>>
+  formatValidationError?(error: unknown): string
+}
+export type DefWithoutID<
+  Parameters extends Schema.Decoder<unknown> = Schema.Decoder<unknown>,
+  M extends Metadata = Metadata,
+> = Omit<Def<Parameters, M>, "id">
+
+export interface Info<
+  Parameters extends Schema.Decoder<unknown> = Schema.Decoder<unknown>,
+  M extends Metadata = Metadata,
+> {
+  id: string
+  init: () => Effect.Effect<DefWithoutID<Parameters, M>>
+}
+
+type Init<Parameters extends Schema.Decoder<unknown>, M extends Metadata> =
+  | DefWithoutID<Parameters, M>
+  | (() => Effect.Effect<DefWithoutID<Parameters, M>>)
+
+export type InferParameters<T> =
+  T extends Info<infer P, any>
+    ? Schema.Schema.Type<P>
+    : T extends Effect.Effect<Info<infer P, any>, any, any>
+      ? Schema.Schema.Type<P>
+      : never
+export type InferMetadata<T> =
+  T extends Info<any, infer M> ? M : T extends Effect.Effect<Info<any, infer M>, any, any> ? M : never
+
+export type InferDef<T> =
+  T extends Info<infer P, infer M>
+    ? Def<P, M>
+    : T extends Effect.Effect<Info<infer P, infer M>, any, any>
+      ? Def<P, M>
+      : never
+
+function wrap<Parameters extends Schema.Decoder<unknown>, Result extends Metadata>(
+  id: string,
+  init: Init<Parameters, Result>,
+  truncate: Truncate.Interface,
+  agents: Agent.Interface,
+) {
+  return () =>
+    Effect.gen(function* () {
+      const toolInfo = typeof init === "function" ? { ...(yield* init()) } : { ...init }
+      // Compile the parser closure once per tool init; `decodeUnknownEffect`
+      // allocates a new closure per call, so hoisting avoids re-closing it for
+      // every LLM tool invocation.
+      const decode = Schema.decodeUnknownEffect(toolInfo.parameters)
+      const execute = toolInfo.execute
+      toolInfo.execute = (args, ctx) => {
+        const attrs = {
+          "tool.name": id,
+          "session.id": ctx.sessionID,
+          "message.id": ctx.messageID,
+          ...(ctx.callID ? { "tool.call_id": ctx.callID } : {}),
+        }
+        return Effect.gen(function* () {
+          const decoded = yield* decode(args).pipe(
+            Effect.mapError((error) =>
+              toolInfo.formatValidationError
+                ? new Error(toolInfo.formatValidationError(error), { cause: error })
+                : new Error(
+                    `The ${id} tool was called with invalid arguments: ${error}.\nPlease rewrite the input so it satisfies the expected schema.`,
+                    { cause: error },
+                  ),
+            ),
+          )
+          const result = yield* execute(decoded as Schema.Schema.Type<Parameters>, ctx)
+          if (result.metadata.truncated !== undefined) {
+            return result
+          }
+          const agent = yield* agents.get(ctx.agent)
+          const truncated = yield* truncate.output(result.output, {}, agent)
+          return {
+            ...result,
+            output: truncated.content,
+            metadata: {
+              ...result.metadata,
+              truncated: truncated.truncated,
+              ...(truncated.truncated && { outputPath: truncated.outputPath }),
+            },
+          }
+        }).pipe(Effect.orDie, Effect.withSpan("Tool.execute", { attributes: attrs }))
+      }
+      return toolInfo
+    })
+}
+
+export function define<
+  Parameters extends Schema.Decoder<unknown>,
+  Result extends Metadata,
+  R,
+  ID extends string = string,
+>(
+  id: ID,
+  init: Effect.Effect<Init<Parameters, Result>, never, R>,
+): Effect.Effect<Info<Parameters, Result>, never, R | Truncate.Service | Agent.Service> & { id: ID } {
+  return Object.assign(
+    Effect.gen(function* () {
+      const resolved = yield* init
+      const truncate = yield* Truncate.Service
+      const agents = yield* Agent.Service
+      return { id, init: wrap(id, resolved, truncate, agents) }
+    }),
+    { id },
+  )
+}
+
+export function init<P extends Schema.Decoder<unknown>, M extends Metadata>(
+  info: Info<P, M>,
+): Effect.Effect<Def<P, M>> {
+  return Effect.gen(function* () {
+    const init = yield* info.init()
+    return {
+      ...init,
+      id: info.id,
+    }
+  })
+}
+
+export * as Tool from "./tool"