langflower 0.1.0 → 0.1.2

package/vendor/tools/dist/create-web-fetch.js

@@ -1,164 +0,0 @@
-import net from 'node:net';
-import { Agent, fetch as undiciFetch } from 'undici';
-import { assertUrlSafeForFetch } from './ssrf-guard.js';
-const DEFAULT_TIMEOUT_MS = 30_000;
-const DEFAULT_MAX_BYTES = 5_000_000;
-const MAX_REDIRECTS = 5;
-const headersToRecord = (headers) => {
-    const out = {};
-    headers.forEach((value, key) => {
-        out[key] = value;
-    });
-    return out;
-};
-const readBodyLimited = async (response, maxBytes) => {
-    if (response.body === null) {
-        return '';
-    }
-    const reader = response.body.getReader();
-    const decoder = new TextDecoder();
-    let total = 0;
-    let text = '';
-    for (;;) {
-        const { done, value } = await reader.read();
-        if (done) {
-            break;
-        }
-        if (value === undefined) {
-            continue;
-        }
-        total += value.byteLength;
-        if (total > maxBytes) {
-            await reader.cancel();
-            throw new Error(`Response exceeded maxBytes (${maxBytes}).`);
-        }
-        text += decoder.decode(value, { stream: true });
-    }
-    text += decoder.decode();
-    return text;
-};
-/**
- * Fetch with TCP connect pinned to a previously SSRF-validated address so
- * DNS cannot rebind between check and connect (Host/SNI still use hostname).
- */
-const fetchPinned = async (url, pinnedAddresses, init) => {
-    const pinned = pinnedAddresses[0];
-    // undici RequestInit ≠ DOM RequestInit under exactOptionalPropertyTypes.
-    const undiciInit = init;
-    if (pinned === undefined) {
-        return undiciFetch(url.toString(), undiciInit);
-    }
-    const family = net.isIP(pinned) === 6 ? 6 : 4;
-    const agent = new Agent({
-        connect: {
-            lookup: (_hostname, _options, callback) => {
-                callback(null, pinned, family);
-            },
-        },
-    });
-    try {
-        return (await undiciFetch(url.toString(), {
-            ...undiciInit,
-            dispatcher: agent,
-        }));
-    }
-    finally {
-        await agent.close();
-    }
-};
-/**
- * SSRF-guarded HTTP GET for crawl / Fetch URL nodes.
- * Validates every redirect hop; truncates body at `maxBytes`.
- * Default path pins DNS via undici Agent; injected `fetchImpl` skips pin
- * (tests) but still runs {@link assertUrlSafeForFetch} per hop.
- */
-export const createWebFetch = (options = {}) => {
-    const fetchImpl = options.fetchImpl;
-    return async (request) => {
-        const timeoutMs = typeof request.timeoutMs === 'number' && request.timeoutMs > 0
-            ? request.timeoutMs
-            : DEFAULT_TIMEOUT_MS;
-        const maxBytes = typeof request.maxBytes === 'number' && request.maxBytes > 0
-            ? request.maxBytes
-            : DEFAULT_MAX_BYTES;
-        try {
-            let currentUrl = request.url;
-            for (let hop = 0; hop <= MAX_REDIRECTS; hop += 1) {
-                const safe = await assertUrlSafeForFetch(currentUrl, {
-                    ...(options.allowedHosts !== undefined
-                        ? { allowedHosts: options.allowedHosts }
-                        : {}),
-                });
-                const safeUrl = safe.url;
-                const controller = new AbortController();
-                const timer = setTimeout(() => controller.abort(), timeoutMs);
-                try {
-                    const init = {
-                        method: 'GET',
-                        redirect: 'manual',
-                        signal: controller.signal,
-                        headers: {
-                            Accept: 'text/html,application/xhtml+xml,*/*;q=0.8',
-                            'User-Agent': 'LangflowerCrawl/0.1',
-                        },
-                    };
-                    const response = fetchImpl !== undefined
-                        ? await fetchImpl(safeUrl.toString(), init)
-                        : await fetchPinned(safeUrl, safe.pinnedAddresses, init);
-                    if (response.status >= 300 &&
-                        response.status < 400 &&
-                        response.headers.has('location')) {
-                        const location = response.headers.get('location');
-                        if (location === null || location.length === 0) {
-                            return {
-                                ok: false,
-                                status: response.status,
-                                url: safeUrl.toString(),
-                                headers: headersToRecord(response.headers),
-                                body: '',
-                                error: 'Redirect missing Location header.',
-                            };
-                        }
-                        currentUrl = new URL(location, safeUrl).toString();
-                        continue;
-                    }
-                    const body = await readBodyLimited(response, maxBytes);
-                    return {
-                        ok: response.ok,
-                        status: response.status,
-                        url: safeUrl.toString(),
-                        headers: headersToRecord(response.headers),
-                        body,
-                        ...(response.ok
-                            ? {}
-                            : {
-                                error: `HTTP ${response.status} ${response.statusText}`,
-                            }),
-                    };
-                }
-                finally {
-                    clearTimeout(timer);
-                }
-            }
-            return {
-                ok: false,
-                status: 0,
-                url: currentUrl,
-                headers: {},
-                body: '',
-                error: `Too many redirects (>${MAX_REDIRECTS}).`,
-            };
-        }
-        catch (error) {
-            const message = error instanceof Error ? error.message : String(error);
-            return {
-                ok: false,
-                status: 0,
-                url: request.url,
-                headers: {},
-                body: '',
-                error: message,
-            };
-        }
-    };
-};

package/vendor/tools/dist/domain/args.d.ts

@@ -1,3 +0,0 @@
-export declare const asString: (args: Readonly<Record<string, unknown>>, key: string) => string | undefined;
-export declare const asNumber: (args: Readonly<Record<string, unknown>>, key: string) => number | undefined;
-export declare const requireString: (args: Readonly<Record<string, unknown>>, key: string) => string;

package/vendor/tools/dist/domain/args.js

@@ -1,22 +0,0 @@
-export const asString = (args, key) => {
-    const value = args[key];
-    return typeof value === 'string' ? value : undefined;
-};
-export const asNumber = (args, key) => {
-    const value = args[key];
-    if (typeof value === 'number' && Number.isFinite(value)) {
-        return value;
-    }
-    if (typeof value === 'string' && value.trim().length > 0) {
-        const parsed = Number(value);
-        return Number.isFinite(parsed) ? parsed : undefined;
-    }
-    return undefined;
-};
-export const requireString = (args, key) => {
-    const value = asString(args, key)?.trim();
-    if (value === undefined || value.length === 0) {
-        throw new Error(`Missing required string argument «${key}».`);
-    }
-    return value;
-};

package/vendor/tools/dist/domain/create-domain-handlers.d.ts

@@ -1,22 +0,0 @@
-import type { CrawlContext } from '../create-crawl-context.js';
-import type { KbContext, KbEmbedTexts } from '../kb/create-kb-context.js';
-import type { MemoryContext } from '../create-memory-context.js';
-import type { WebFetchRequest, WebFetchResult } from '../create-web-fetch.js';
-export type DomainToolHandler = (args: Readonly<Record<string, unknown>>) => Promise<string>;
-export type DomainHandlerDeps = {
-    readonly webFetch: (request: WebFetchRequest) => Promise<WebFetchResult>;
-    readonly crawl: CrawlContext;
-    readonly kb: KbContext;
-    readonly memory: MemoryContext;
-    readonly embedTexts?: KbEmbedTexts;
-};
-declare const DOMAIN_TOOL_IDS: readonly ["crawl_fetch", "crawl_extract_links", "crawl_save_page", "crawl_bfs", "kb_ingest", "kb_embed", "kb_search", "kb_list", "kb_delete", "memory_get", "memory_set", "memory_list", "memory_delete"];
-export type DomainToolId = (typeof DOMAIN_TOOL_IDS)[number];
-export declare const isDomainToolId: (toolId: string) => toolId is DomainToolId;
-/**
- * Invokable crawl / KB / Memory handlers for `createProjectHarness`.
- * Not listed by `listBuiltinRegistrations` — inventory only when wired.
- */
-export declare const createDomainHandlers: (deps: DomainHandlerDeps) => Readonly<Record<string, DomainToolHandler>>;
-export {};
-//# sourceMappingURL=create-domain-handlers.d.ts.map

package/vendor/tools/dist/domain/create-domain-handlers.js

@@ -1,223 +0,0 @@
-import { asNumber, asString, requireString } from './args.js';
-import { extractHtmlTitle, extractLinks, htmlToText, isSameHost, } from './html.js';
-const DOMAIN_TOOL_IDS = [
-    'crawl_fetch',
-    'crawl_extract_links',
-    'crawl_save_page',
-    'crawl_bfs',
-    'kb_ingest',
-    'kb_embed',
-    'kb_search',
-    'kb_list',
-    'kb_delete',
-    'memory_get',
-    'memory_set',
-    'memory_list',
-    'memory_delete',
-];
-export const isDomainToolId = (toolId) => DOMAIN_TOOL_IDS.includes(toolId);
-const json = (value) => `${JSON.stringify(value, null, 2)}\n`;
-const fetchPage = async (webFetch, url, timeoutMs, maxBytes) => {
-    const result = await webFetch({
-        url,
-        ...(timeoutMs !== undefined ? { timeoutMs } : {}),
-        ...(maxBytes !== undefined ? { maxBytes } : {}),
-    });
-    if (!result.ok) {
-        throw new Error(result.error ?? `Fetch failed for ${url} (status ${result.status}).`);
-    }
-    const text = htmlToText(result.body);
-    const title = extractHtmlTitle(result.body);
-    return {
-        url: result.url,
-        html: result.body,
-        text,
-        ...(title !== undefined ? { title } : {}),
-        status: result.status,
-    };
-};
-/**
- * Invokable crawl / KB / Memory handlers for `createProjectHarness`.
- * Not listed by `listBuiltinRegistrations` — inventory only when wired.
- */
-export const createDomainHandlers = (deps) => {
-    const crawl_fetch = async (args) => {
-        const page = await fetchPage(deps.webFetch, requireString(args, 'url'), asNumber(args, 'timeoutMs'), asNumber(args, 'maxBytes'));
-        return json({
-            url: page.url,
-            status: page.status,
-            title: page.title ?? null,
-            text: page.text,
-            htmlChars: page.html.length,
-        });
-    };
-    const crawl_extract_links = async (args) => {
-        const html = requireString(args, 'html');
-        const baseUrl = requireString(args, 'baseUrl');
-        const links = extractLinks(html, baseUrl);
-        return json({ count: links.length, links });
-    };
-    const crawl_save_page = async (args) => {
-        const url = requireString(args, 'url');
-        const html = asString(args, 'html') ?? '';
-        const text = asString(args, 'text') ?? htmlToText(html);
-        const title = asString(args, 'title');
-        const saved = await deps.crawl.savePage({
-            url,
-            html,
-            text,
-            ...(title !== undefined && title.length > 0 ? { title } : {}),
-        });
-        return json(saved);
-    };
-    const crawl_bfs = async (args) => {
-        const startUrl = requireString(args, 'url');
-        const maxPages = Math.min(Math.max(1, Math.floor(asNumber(args, 'maxPages') ?? 5)), 50);
-        const sameHostOnly = args.sameHostOnly !== false;
-        const save = args.save !== false;
-        const queue = [startUrl];
-        const seen = new Set();
-        const pages = [];
-        while (queue.length > 0 && pages.length < maxPages) {
-            const url = queue.shift();
-            if (url === undefined || seen.has(url)) {
-                continue;
-            }
-            seen.add(url);
-            try {
-                const page = await fetchPage(deps.webFetch, url);
-                if (save) {
-                    await deps.crawl.savePage({
-                        url: page.url,
-                        html: page.html,
-                        text: page.text,
-                        ...(page.title !== undefined
-                            ? { title: page.title }
-                            : {}),
-                    });
-                }
-                pages.push({
-                    url: page.url,
-                    title: page.title ?? null,
-                    textChars: page.text.length,
-                });
-                for (const link of extractLinks(page.html, page.url)) {
-                    if (seen.has(link) || queue.includes(link)) {
-                        continue;
-                    }
-                    if (sameHostOnly && !isSameHost(startUrl, link)) {
-                        continue;
-                    }
-                    queue.push(link);
-                }
-            }
-            catch (error) {
-                pages.push({
-                    url,
-                    error: error instanceof Error
-                        ? error.message
-                        : String(error),
-                });
-            }
-        }
-        return json({
-            startUrl,
-            visited: pages.length,
-            pages,
-        });
-    };
-    const kb_ingest = async (args) => {
-        const documentId = asString(args, 'documentId')?.trim();
-        const source = asString(args, 'source')?.trim();
-        const chunkIds = await deps.kb.ingest({
-            collectionId: requireString(args, 'collectionId'),
-            text: requireString(args, 'text'),
-            ...(documentId !== undefined && documentId.length > 0
-                ? { documentId }
-                : {}),
-            ...(source !== undefined && source.length > 0 ? { source } : {}),
-        });
-        return json({ chunkIds, count: chunkIds.length });
-    };
-    const kb_embed = async (args) => {
-        const count = await deps.kb.embed({
-            collectionId: requireString(args, 'collectionId'),
-            ...(deps.embedTexts !== undefined
-                ? { embedTexts: deps.embedTexts }
-                : {}),
-        });
-        return json({ embedded: count });
-    };
-    const kb_search = async (args) => {
-        const topK = asNumber(args, 'topK');
-        const hits = await deps.kb.search({
-            collectionId: requireString(args, 'collectionId'),
-            query: requireString(args, 'query'),
-            ...(topK !== undefined ? { topK } : {}),
-            ...(deps.embedTexts !== undefined
-                ? { embedTexts: deps.embedTexts }
-                : {}),
-        });
-        return json({ hits });
-    };
-    const kb_list = async (args) => {
-        const collectionId = asString(args, 'collectionId')?.trim();
-        const items = await deps.kb.list({
-            ...(collectionId !== undefined && collectionId.length > 0
-                ? { collectionId }
-                : {}),
-        });
-        return json({ items });
-    };
-    const kb_delete = async (args) => {
-        const chunkIdsRaw = args.chunkIds;
-        const chunkIds = Array.isArray(chunkIdsRaw)
-            ? chunkIdsRaw.map(String)
-            : undefined;
-        const documentId = asString(args, 'documentId')?.trim();
-        const deleted = await deps.kb.delete({
-            collectionId: requireString(args, 'collectionId'),
-            ...(documentId !== undefined && documentId.length > 0
-                ? { documentId }
-                : {}),
-            ...(chunkIds !== undefined ? { chunkIds } : {}),
-        });
-        return json({ deleted });
-    };
-    const memory_get = async (args) => {
-        const key = requireString(args, 'key');
-        const value = await deps.memory.get(key);
-        return json({ key, value: value ?? null });
-    };
-    const memory_set = async (args) => {
-        const key = requireString(args, 'key');
-        const value = requireString(args, 'value');
-        await deps.memory.set(key, value);
-        return json({ key, ok: true });
-    };
-    const memory_list = async () => {
-        const keys = await deps.memory.list();
-        return json({ keys });
-    };
-    const memory_delete = async (args) => {
-        const key = requireString(args, 'key');
-        const deleted = await deps.memory.delete(key);
-        return json({ key, deleted });
-    };
-    return {
-        crawl_fetch,
-        crawl_extract_links,
-        crawl_save_page,
-        crawl_bfs,
-        kb_ingest,
-        kb_embed,
-        kb_search,
-        kb_list,
-        kb_delete,
-        memory_get,
-        memory_set,
-        memory_list,
-        memory_delete,
-    };
-};
-//# sourceMappingURL=create-domain-handlers.js.map

package/vendor/tools/dist/domain/domain-tool-configs.d.ts

@@ -1,30 +0,0 @@
-import type { WebFetchRequest, WebFetchResult } from '../create-web-fetch.js';
-/**
- * Invoke-time context for domain tool handlers.
- * Extends the SDK identity `{ projectDir, runId }` with optional host hooks
- * (authorize / fetch / path policy) supplied by the agent shell —
- * not part of the author `@langflower/node-sdk` surface.
- */
-export type ToolHandlerContext = {
-    readonly projectDir: string;
-    readonly runId: string;
-    readonly authorize?: (call: {
-        readonly toolId: string;
-        readonly args: Readonly<Record<string, unknown>>;
-    }) => Promise<'allow' | 'deny'>;
-    readonly webFetch?: (request: WebFetchRequest) => Promise<WebFetchResult>;
-    readonly denyPaths?: readonly string[];
-    readonly allowedHosts?: readonly string[];
-    /** Per-invoke abort (tool timeout / run cancel) for builtins. */
-    readonly signal?: AbortSignal;
-};
-export type ToolHandler = (args: Readonly<Record<string, unknown>>, ctx: ToolHandlerContext) => Promise<string>;
-export type DomainToolConfig = {
-    readonly toolId: string;
-    readonly name?: string;
-    readonly description: string;
-    readonly inputSchema: object;
-    readonly handler: ToolHandler;
-};
-export declare const CRAWL_TOOL_CONFIGS: readonly DomainToolConfig[];
-export declare const MEMORY_TOOL_CONFIGS: readonly DomainToolConfig[];