afterburn-cli 1.0.0

package/dist/discovery/element-mapper.js

@@ -0,0 +1,422 @@
+// Interactive element discovery including hidden elements (modals, dropdowns, mobile nav)
+// Destructive action denylist - skip elements with these keywords
+const DESTRUCTIVE_KEYWORDS = [
+    'delete',
+    'remove',
+    'destroy',
+    'reset',
+    'clear',
+    'drop',
+    'purge',
+    'revoke',
+    'terminate',
+    'unsubscribe',
+    'cancel-account',
+    'close-account',
+    'logout',
+    'log out',
+    'sign out',
+];
+/**
+ * Check if text matches destructive action patterns
+ */
+function isDestructiveAction(text) {
+    const lowerText = text.toLowerCase().trim();
+    return DESTRUCTIVE_KEYWORDS.some(keyword => lowerText.includes(keyword));
+}
+function escapeSelectorText(value) {
+    return value.replace(/\\/g, '\\\\').replace(/"/g, '\\"');
+}
+/**
+ * Discover all visible interactive elements on a page
+ */
+export async function discoverElements(page, pageUrl) {
+    const elements = {
+        forms: [],
+        buttons: [],
+        links: [],
+        menus: [],
+        otherInteractive: [],
+    };
+    const pageHostname = new URL(pageUrl).hostname;
+    // Discover all forms with field inventory
+    const forms = await page.locator('form').all();
+    for (let i = 0; i < forms.length; i++) {
+        const form = forms[i];
+        try {
+            const action = (await form.getAttribute('action')) || '';
+            const method = ((await form.getAttribute('method')) || 'GET').toUpperCase();
+            // Build selector for the form
+            const formId = await form.getAttribute('id');
+            const formName = await form.getAttribute('name');
+            let selector = 'form';
+            if (formId) {
+                selector = `form#${formId}`;
+            }
+            else if (formName) {
+                selector = `form[name="${formName}"]`;
+            }
+            else {
+                selector = `form:nth-of-type(${i + 1})`;
+            }
+            // Resolve action URL
+            const resolvedAction = action ? new URL(action, pageUrl).href : pageUrl;
+            // Find all input, textarea, select within the form
+            const fields = await form.locator('input, textarea, select').all();
+            const formFields = [];
+            for (const field of fields) {
+                try {
+                    const type = (await field.getAttribute('type')) || 'text';
+                    const name = (await field.getAttribute('name')) || '';
+                    const required = (await field.getAttribute('required')) !== null;
+                    const placeholder = (await field.getAttribute('placeholder')) || '';
+                    // Try to find associated label
+                    let label = (await field.getAttribute('aria-label')) || '';
+                    if (!label) {
+                        const fieldId = await field.getAttribute('id');
+                        if (fieldId) {
+                            // Look for label with matching "for" attribute
+                            const labelElement = await page.locator(`label[for="${fieldId}"]`).first();
+                            const labelText = await labelElement.textContent().catch(() => '');
+                            label = labelText?.trim() || '';
+                        }
+                    }
+                    // If still no label, check if field is inside a label
+                    if (!label) {
+                        const parentLabel = await field.locator('xpath=ancestor::label[1]').first();
+                        const labelText = await parentLabel.textContent().catch(() => '');
+                        label = labelText?.trim() || '';
+                    }
+                    formFields.push({
+                        type,
+                        name,
+                        label,
+                        required,
+                        placeholder,
+                    });
+                }
+                catch {
+                    // Failed to extract field info, skip
+                }
+            }
+            elements.forms.push({
+                action: resolvedAction,
+                method,
+                selector,
+                fields: formFields,
+            });
+        }
+        catch {
+            // Failed to extract form info, skip
+        }
+    }
+    // Discover all buttons
+    const buttons = await page.getByRole('button').all();
+    for (let i = 0; i < buttons.length; i++) {
+        const button = buttons[i];
+        try {
+            const text = (await button.textContent()) || '';
+            const ariaLabel = (await button.getAttribute('aria-label')) || '';
+            const type = (await button.getAttribute('type')) || 'button';
+            const disabled = (await button.getAttribute('disabled')) !== null;
+            const isVisible = await button.isVisible();
+            // Build selector - prefer has-text when text is unique
+            let selector = 'button';
+            if (text.trim()) {
+                selector = `button:has-text("${escapeSelectorText(text.trim())}")`;
+            }
+            else if (ariaLabel) {
+                selector = `button[aria-label="${escapeSelectorText(ariaLabel)}"]`;
+            }
+            else {
+                // Fallback to nth-of-type
+                selector = `button:nth-of-type(${i + 1})`;
+            }
+            elements.buttons.push({
+                type: 'button',
+                selector,
+                text: text.trim() || ariaLabel,
+                visible: isVisible,
+                attributes: {
+                    type,
+                    disabled: disabled.toString(),
+                },
+            });
+        }
+        catch {
+            // Failed to extract button info, skip
+        }
+    }
+    // Discover all links with href
+    const links = await page.locator('a[href]').all();
+    const seenHrefs = new Set();
+    for (const link of links) {
+        try {
+            const href = await link.getAttribute('href');
+            if (!href)
+                continue;
+            // Resolve to absolute URL
+            const absoluteUrl = new URL(href, pageUrl).href;
+            // Deduplicate by href
+            if (seenHrefs.has(absoluteUrl))
+                continue;
+            seenHrefs.add(absoluteUrl);
+            const text = (await link.textContent())?.trim() || '';
+            // Determine if internal
+            const linkHostname = new URL(absoluteUrl).hostname;
+            const isInternal = linkHostname === pageHostname;
+            elements.links.push({
+                href: absoluteUrl,
+                text,
+                isInternal,
+            });
+        }
+        catch {
+            // Failed to extract link info or invalid URL, skip
+        }
+    }
+    // Discover navigation menus
+    const navContainers = await page
+        .locator('nav, [role="navigation"], [role="menubar"]')
+        .all();
+    for (let i = 0; i < navContainers.length; i++) {
+        const nav = navContainers[i];
+        try {
+            const text = (await nav.textContent())?.trim() || '';
+            const ariaLabel = (await nav.getAttribute('aria-label')) || '';
+            const isVisible = await nav.isVisible();
+            elements.menus.push({
+                type: 'menu',
+                selector: `nav:nth-of-type(${i + 1})`,
+                text: ariaLabel || text.substring(0, 50) || `Menu ${i + 1}`,
+                visible: isVisible,
+                attributes: {},
+            });
+        }
+        catch {
+            // Failed to extract menu info, skip
+        }
+    }
+    // Discover other interactive elements
+    const interactiveRoles = ['tab', 'tabpanel', 'dialog', 'combobox', 'listbox'];
+    for (const role of interactiveRoles) {
+        const roleElements = await page.locator(`[role="${role}"]`).all();
+        for (let i = 0; i < roleElements.length; i++) {
+            const element = roleElements[i];
+            try {
+                const text = (await element.textContent())?.trim() || '';
+                const ariaLabel = (await element.getAttribute('aria-label')) || '';
+                const isVisible = await element.isVisible();
+                elements.otherInteractive.push({
+                    type: role, // TypeScript will coerce to InteractiveElement['type']
+                    selector: `[role="${role}"]:nth-of-type(${i + 1})`,
+                    text: ariaLabel || text.substring(0, 50) || `${role} ${i + 1}`,
+                    visible: isVisible,
+                    attributes: { role },
+                });
+            }
+            catch {
+                // Failed to extract element info, skip
+            }
+        }
+    }
+    return elements;
+}
+/**
+ * Discover hidden elements by triggering interactions (buttons, menus, modals)
+ * Returns only NEWLY discovered elements (not present in initial page load)
+ */
+export async function discoverHiddenElements(page, pageUrl) {
+    // Take "before" snapshot of visible element count
+    const beforeCounts = {
+        buttons: (await page.getByRole('button').all()).length,
+        links: (await page.locator('a[href]').all()).length,
+        forms: (await page.locator('form').all()).length,
+    };
+    const newElements = {
+        forms: [],
+        buttons: [],
+        links: [],
+        menus: [],
+        otherInteractive: [],
+    };
+    // Find buttons that might trigger hidden content
+    const triggerSelectors = [
+        'button[aria-expanded]',
+        'button[aria-haspopup]',
+        'button[data-toggle]',
+        'button[data-bs-toggle]',
+        '[class*="hamburger"]',
+        '[class*="menu-toggle"]',
+        '[class*="navbar-toggler"]',
+    ];
+    // Also find buttons with trigger-like text
+    const triggerTexts = ['menu', 'nav', 'more', 'show', 'open', 'expand'];
+    const allButtons = await page.getByRole('button').all();
+    const triggerButtons = [];
+    // Collect explicit trigger buttons
+    for (const selector of triggerSelectors) {
+        const buttons = await page.locator(selector).all();
+        triggerButtons.push(...buttons);
+    }
+    // Collect buttons with trigger-like text
+    for (const button of allButtons) {
+        try {
+            const text = ((await button.textContent()) || '').toLowerCase();
+            if (triggerTexts.some((word) => text.includes(word))) {
+                triggerButtons.push(button);
+            }
+        }
+        catch {
+            // Skip
+        }
+    }
+    // Click each trigger button and check for new elements (cap at 10 to avoid perf explosion)
+    for (const trigger of triggerButtons.slice(0, 10)) {
+        try {
+            // Check if button is visible and enabled
+            const isVisible = await trigger.isVisible();
+            if (!isVisible)
+                continue;
+            // Check for destructive actions
+            const buttonText = (await trigger.textContent()) || '';
+            const ariaLabel = (await trigger.getAttribute('aria-label')) || '';
+            const combinedText = `${buttonText} ${ariaLabel}`;
+            if (isDestructiveAction(combinedText)) {
+                continue; // Skip destructive buttons
+            }
+            // Click the trigger
+            await trigger.click({ timeout: 2000 });
+            // Wait for animations
+            await page.waitForTimeout(500);
+            // Check for newly visible elements
+            const modalSelectors = [
+                '[role="dialog"]',
+                '[role="menu"]',
+                '.modal',
+                '.dropdown-menu',
+                '.nav-menu',
+                '[aria-hidden="false"]',
+            ];
+            let foundNewContent = false;
+            for (const selector of modalSelectors) {
+                const modalElements = await page.locator(selector).all();
+                for (const modal of modalElements) {
+                    try {
+                        const isVisible = await modal.isVisible();
+                        if (isVisible) {
+                            foundNewContent = true;
+                            // Run discoverElements on the new content
+                            const modalElements = await discoverElements(page, pageUrl);
+                            // Merge with newElements (deduplicate by selector)
+                            for (const form of modalElements.forms) {
+                                if (!newElements.forms.some((f) => f.selector === form.selector)) {
+                                    newElements.forms.push(form);
+                                }
+                            }
+                            for (const button of modalElements.buttons) {
+                                if (!newElements.buttons.some((b) => b.selector === button.selector)) {
+                                    newElements.buttons.push(button);
+                                }
+                            }
+                            for (const link of modalElements.links) {
+                                if (!newElements.links.some((l) => l.href === link.href)) {
+                                    newElements.links.push(link);
+                                }
+                            }
+                            for (const menu of modalElements.menus) {
+                                if (!newElements.menus.some((m) => m.selector === menu.selector)) {
+                                    newElements.menus.push(menu);
+                                }
+                            }
+                            for (const other of modalElements.otherInteractive) {
+                                if (!newElements.otherInteractive.some((o) => o.selector === other.selector)) {
+                                    newElements.otherInteractive.push(other);
+                                }
+                            }
+                        }
+                    }
+                    catch {
+                        // Skip
+                    }
+                }
+            }
+            // Try to close/reset
+            if (foundNewContent) {
+                try {
+                    // Try clicking trigger again to toggle off
+                    await trigger.click({ timeout: 1000 });
+                    await page.waitForTimeout(300);
+                }
+                catch {
+                    // Try pressing Escape
+                    try {
+                        await page.keyboard.press('Escape');
+                        await page.waitForTimeout(300);
+                    }
+                    catch {
+                        // If can't close, reload the page to reset state
+                        try {
+                            await page.reload({ timeout: 5000, waitUntil: 'domcontentloaded' });
+                        }
+                        catch {
+                            // Can't reload, break loop
+                            break;
+                        }
+                    }
+                }
+            }
+        }
+        catch {
+            // Failed to interact with trigger, continue
+            continue;
+        }
+    }
+    // Also try hovering on nav items to discover dropdowns
+    const navItems = await page.locator('nav li, [role="menubar"] > *').all();
+    for (const item of navItems) {
+        try {
+            const isVisible = await item.isVisible();
+            if (!isVisible)
+                continue;
+            // Hover on the item
+            await item.hover({ timeout: 1000 });
+            await page.waitForTimeout(300);
+            // Check for dropdown menus
+            const dropdowns = await page.locator('.dropdown-menu, [role="menu"]').all();
+            for (const dropdown of dropdowns) {
+                try {
+                    const isDropdownVisible = await dropdown.isVisible();
+                    if (isDropdownVisible) {
+                        // Discover elements in dropdown
+                        const dropdownElements = await discoverElements(page, pageUrl);
+                        // Merge new links
+                        for (const link of dropdownElements.links) {
+                            if (!newElements.links.some((l) => l.href === link.href)) {
+                                newElements.links.push(link);
+                            }
+                        }
+                        // Merge new buttons
+                        for (const button of dropdownElements.buttons) {
+                            if (!newElements.buttons.some((b) => b.selector === button.selector)) {
+                                newElements.buttons.push(button);
+                            }
+                        }
+                    }
+                }
+                catch {
+                    // Skip
+                }
+            }
+        }
+        catch {
+            // Failed to hover, continue
+            continue;
+        }
+    }
+    // Filter to return only NEWLY discovered elements (subtract "before" snapshot)
+    // Since we already deduplicated during collection, and we're discovering from hidden content,
+    // most of these should be new. But let's do a final pass to be safe.
+    return newElements;
+}
+//# sourceMappingURL=element-mapper.js.map

package/dist/discovery/element-mapper.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"element-mapper.js","sourceRoot":"","sources":["../../src/discovery/element-mapper.ts"],"names":[],"mappings":"AAAA,0FAA0F;AAK1F,kEAAkE;AAClE,MAAM,oBAAoB,GAAG;IAC3B,QAAQ;IACR,QAAQ;IACR,SAAS;IACT,OAAO;IACP,OAAO;IACP,MAAM;IACN,OAAO;IACP,QAAQ;IACR,WAAW;IACX,aAAa;IACb,gBAAgB;IAChB,eAAe;IACf,QAAQ;IACR,SAAS;IACT,UAAU;CACX,CAAC;AAEF;;GAEG;AACH,SAAS,mBAAmB,CAAC,IAAY;IACvC,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC;IAC5C,OAAO,oBAAoB,CAAC,IAAI,CAAC,OAAO,CAAC,EAAE,CAAC,SAAS,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC;AAC3E,CAAC;AAED,SAAS,kBAAkB,CAAC,KAAa;IACvC,OAAO,KAAK,CAAC,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,OAAO,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;AAC3D,CAAC;AAaD;;GAEG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CAAC,IAAU,EAAE,OAAe;IAChE,MAAM,QAAQ,GAAuB;QACnC,KAAK,EAAE,EAAE;QACT,OAAO,EAAE,EAAE;QACX,KAAK,EAAE,EAAE;QACT,KAAK,EAAE,EAAE;QACT,gBAAgB,EAAE,EAAE;KACrB,CAAC;IAEF,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,CAAC,QAAQ,CAAC;IAE/C,0CAA0C;IAC1C,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC;IAC/C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACtC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QAEtB,IAAI,CAAC;YACH,MAAM,MAAM,GAAG,CAAC,MAAM,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAC;YACzD,MAAM,MAAM,GAAG,CAAC,CAAC,MAAM,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC;YAE5E,8BAA8B;YAC9B,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;YAC7C,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC;YACjD,IAAI,QAAQ,GAAG,MAAM,CAAC;YACtB,IAAI,MAAM,EAAE,CAAC;gBACX,QAAQ,GAAG,QAAQ,MAAM,EAAE,CAAC;YAC9B,CAAC;iBAAM,IAAI,QAAQ,EAAE,CAAC;gBACpB,QAAQ,GAAG,cAAc,QAAQ,IAAI,CAAC;YACxC,CAAC;iBAAM,CAAC;gBACN,QAAQ,GAAG,oBAAoB,CAAC,GAAG,CAAC,GAAG,CAAC;YAC1C,CAAC;YAED,qBAAqB;YACrB,MAAM,cAAc,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC;YAExE,mDAAmD;YACnD,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,yBAAyB,CAAC,CAAC,GAAG,EAAE,CAAC;YACnE,MAAM,UAAU,GAAgB,EAAE,CAAC;YAEnC,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBAC3B,IAAI,CAAC;oBACH,MAAM,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,IAAI,MAAM,CAAC;oBAC1D,MAAM,IAAI,GAAG,CAAC,MAAM,KAAK,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;oBACtD,MAAM,QAAQ,GAAG,CAAC,MAAM,KAAK,CAAC,YAAY,CAAC,UAAU,CAAC,CAAC,KAAK,IAAI,CAAC;oBACjE,MAAM,WAAW,GAAG,CAAC,MAAM,KAAK,CAAC,YAAY,CAAC,aAAa,CAAC,CAAC,IAAI,EAAE,CAAC;oBAEpE,+BAA+B;oBAC/B,IAAI,KAAK,GAAG,CAAC,MAAM,KAAK,CAAC,YAAY,CAAC,YAAY,CAAC,CAAC,IAAI,EAAE,CAAC;oBAC3D,IAAI,CAAC,KAAK,EAAE,CAAC;wBACX,MAAM,OAAO,GAAG,MAAM,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;wBAC/C,IAAI,OAAO,EAAE,CAAC;4BACZ,+CAA+C;4BAC/C,MAAM,YAAY,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,cAAc,OAAO,IAAI,CAAC,CAAC,KAAK,EAAE,CAAC;4BAC3E,MAAM,SAAS,GAAG,MAAM,YAAY,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC;4BACnE,KAAK,GAAG,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;wBAClC,CAAC;oBACH,CAAC;oBAED,sDAAsD;oBACtD,IAAI,CAAC,KAAK,EAAE,CAAC;wBACX,MAAM,WAAW,GAAG,MAAM,KAAK,CAAC,OAAO,CAAC,0BAA0B,CAAC,CAAC,KAAK,EAAE,CAAC;wBAC5E,MAAM,SAAS,GAAG,MAAM,WAAW,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC;wBAClE,KAAK,GAAG,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;oBAClC,CAAC;oBAED,UAAU,CAAC,IAAI,CAAC;wBACd,IAAI;wBACJ,IAAI;wBACJ,KAAK;wBACL,QAAQ;wBACR,WAAW;qBACZ,CAAC,CAAC;gBACL,CAAC;gBAAC,MAAM,CAAC;oBACP,qCAAqC;gBACvC,CAAC;YACH,CAAC;YAED,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC;gBAClB,MAAM,EAAE,cAAc;gBACtB,MAAM;gBACN,QAAQ;gBACR,MAAM,EAAE,UAAU;aACnB,CAAC,CAAC;QACL,CAAC;QAAC,MAAM,CAAC;YACP,oCAAoC;QACtC,CAAC;IACH,CAAC;IAED,uBAAuB;IACvB,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC;IACrD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,MAAM,MAAM,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QAE1B,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,CAAC,MAAM,MAAM,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC;YAChD,MAAM,SAAS,GAAG,CAAC,MAAM,MAAM,CAAC,YAAY,CAAC,YAAY,CAAC,CAAC,IAAI,EAAE,CAAC;YAClE,MAAM,IAAI,GAAG,CAAC,MAAM,MAAM,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,IAAI,QAAQ,CAAC;YAC7D,MAAM,QAAQ,GAAG,CAAC,MAAM,MAAM,CAAC,YAAY,CAAC,UAAU,CAAC,CAAC,KAAK,IAAI,CAAC;YAClE,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,SAAS,EAAE,CAAC;YAE3C,uDAAuD;YACvD,IAAI,QAAQ,GAAG,QAAQ,CAAC;YACxB,IAAI,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;gBAChB,QAAQ,GAAG,oBAAoB,kBAAkB,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,IAAI,CAAC;YACrE,CAAC;iBAAM,IAAI,SAAS,EAAE,CAAC;gBACrB,QAAQ,GAAG,sBAAsB,kBAAkB,CAAC,SAAS,CAAC,IAAI,CAAC;YACrE,CAAC;iBAAM,CAAC;gBACN,0BAA0B;gBAC1B,QAAQ,GAAG,sBAAsB,CAAC,GAAG,CAAC,GAAG,CAAC;YAC5C,CAAC;YAED,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC;gBACpB,IAAI,EAAE,QAAQ;gBACd,QAAQ;gBACR,IAAI,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,SAAS;gBAC9B,OAAO,EAAE,SAAS;gBAClB,UAAU,EAAE;oBACV,IAAI;oBACJ,QAAQ,EAAE,QAAQ,CAAC,QAAQ,EAAE;iBAC9B;aACF,CAAC,CAAC;QACL,CAAC;QAAC,MAAM,CAAC;YACP,sCAAsC;QACxC,CAAC;IACH,CAAC;IAED,+BAA+B;IAC/B,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,GAAG,EAAE,CAAC;IAClD,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;IAEpC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC;YAC7C,IAAI,CAAC,IAAI;gBAAE,SAAS;YAEpB,0BAA0B;YAC1B,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,IAAI,CAAC;YAEhD,sBAAsB;YACtB,IAAI,SAAS,CAAC,GAAG,CAAC,WAAW,CAAC;gBAAE,SAAS;YACzC,SAAS,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;YAE3B,MAAM,IAAI,GAAG,CAAC,MAAM,IAAI,CAAC,WAAW,EAAE,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;YAEtD,wBAAwB;YACxB,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,WAAW,CAAC,CAAC,QAAQ,CAAC;YACnD,MAAM,UAAU,GAAG,YAAY,KAAK,YAAY,CAAC;YAEjD,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC;gBAClB,IAAI,EAAE,WAAW;gBACjB,IAAI;gBACJ,UAAU;aACX,CAAC,CAAC;QACL,CAAC;QAAC,MAAM,CAAC;YACP,mDAAmD;QACrD,CAAC;IACH,CAAC;IAED,4BAA4B;IAC5B,MAAM,aAAa,GAAG,MAAM,IAAI;SAC7B,OAAO,CAAC,4CAA4C,CAAC;SACrD,GAAG,EAAE,CAAC;IAET,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,aAAa,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,GAAG,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;QAE7B,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,WAAW,EAAE,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;YACrD,MAAM,SAAS,GAAG,CAAC,MAAM,GAAG,CAAC,YAAY,CAAC,YAAY,CAAC,CAAC,IAAI,EAAE,CAAC;YAC/D,MAAM,SAAS,GAAG,MAAM,GAAG,CAAC,SAAS,EAAE,CAAC;YAExC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC;gBAClB,IAAI,EAAE,MAAM;gBACZ,QAAQ,EAAE,mBAAmB,CAAC,GAAG,CAAC,GAAG;gBACrC,IAAI,EAAE,SAAS,IAAI,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,QAAQ,CAAC,GAAG,CAAC,EAAE;gBAC3D,OAAO,EAAE,SAAS;gBAClB,UAAU,EAAE,EAAE;aACf,CAAC,CAAC;QACL,CAAC;QAAC,MAAM,CAAC;YACP,oCAAoC;QACtC,CAAC;IACH,CAAC;IAED,sCAAsC;IACtC,MAAM,gBAAgB,GAAG,CAAC,KAAK,EAAE,UAAU,EAAE,QAAQ,EAAE,UAAU,EAAE,SAAS,CAAC,CAAC;IAE9E,KAAK,MAAM,IAAI,IAAI,gBAAgB,EAAE,CAAC;QACpC,MAAM,YAAY,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,UAAU,IAAI,IAAI,CAAC,CAAC,GAAG,EAAE,CAAC;QAElE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC7C,MAAM,OAAO,GAAG,YAAY,CAAC,CAAC,CAAC,CAAC;YAEhC,IAAI,CAAC;gBACH,MAAM,IAAI,GAAG,CAAC,MAAM,OAAO,CAAC,WAAW,EAAE,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;gBACzD,MAAM,SAAS,GAAG,CAAC,MAAM,OAAO,CAAC,YAAY,CAAC,YAAY,CAAC,CAAC,IAAI,EAAE,CAAC;gBACnE,MAAM,SAAS,GAAG,MAAM,OAAO,CAAC,SAAS,EAAE,CAAC;gBAE5C,QAAQ,CAAC,gBAAgB,CAAC,IAAI,CAAC;oBAC7B,IAAI,EAAE,IAAW,EAAE,uDAAuD;oBAC1E,QAAQ,EAAE,UAAU,IAAI,kBAAkB,CAAC,GAAG,CAAC,GAAG;oBAClD,IAAI,EAAE,SAAS,IAAI,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,EAAE;oBAC9D,OAAO,EAAE,SAAS;oBAClB,UAAU,EAAE,EAAE,IAAI,EAAE;iBACrB,CAAC,CAAC;YACL,CAAC;YAAC,MAAM,CAAC;gBACP,uCAAuC;YACzC,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,sBAAsB,CAC1C,IAAU,EACV,OAAe;IAEf,kDAAkD;IAClD,MAAM,YAAY,GAAG;QACnB,OAAO,EAAE,CAAC,MAAM,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,MAAM;QACtD,KAAK,EAAE,CAAC,MAAM,IAAI,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,MAAM;QACnD,KAAK,EAAE,CAAC,MAAM,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,MAAM;KACjD,CAAC;IAEF,MAAM,WAAW,GAAuB;QACtC,KAAK,EAAE,EAAE;QACT,OAAO,EAAE,EAAE;QACX,KAAK,EAAE,EAAE;QACT,KAAK,EAAE,EAAE;QACT,gBAAgB,EAAE,EAAE;KACrB,CAAC;IAEF,iDAAiD;IACjD,MAAM,gBAAgB,GAAG;QACvB,uBAAuB;QACvB,uBAAuB;QACvB,qBAAqB;QACrB,wBAAwB;QACxB,sBAAsB;QACtB,wBAAwB;QACxB,2BAA2B;KAC5B,CAAC;IAEF,2CAA2C;IAC3C,MAAM,YAAY,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC;IAEvE,MAAM,UAAU,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC;IACxD,MAAM,cAAc,GAAG,EAAE,CAAC;IAE1B,mCAAmC;IACnC,KAAK,MAAM,QAAQ,IAAI,gBAAgB,EAAE,CAAC;QACxC,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC;QACnD,cAAc,CAAC,IAAI,CAAC,GAAG,OAAO,CAAC,CAAC;IAClC,CAAC;IAED,yCAAyC;IACzC,KAAK,MAAM,MAAM,IAAI,UAAU,EAAE,CAAC;QAChC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,CAAC,CAAC,MAAM,MAAM,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC;YAChE,IAAI,YAAY,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;gBACrD,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAC9B,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,OAAO;QACT,CAAC;IACH,CAAC;IAED,2FAA2F;IAC3F,KAAK,MAAM,OAAO,IAAI,cAAc,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC;QAClD,IAAI,CAAC;YACH,yCAAyC;YACzC,MAAM,SAAS,GAAG,MAAM,OAAO,CAAC,SAAS,EAAE,CAAC;YAC5C,IAAI,CAAC,SAAS;gBAAE,SAAS;YAEzB,gCAAgC;YAChC,MAAM,UAAU,GAAG,CAAC,MAAM,OAAO,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC;YACvD,MAAM,SAAS,GAAG,CAAC,MAAM,OAAO,CAAC,YAAY,CAAC,YAAY,CAAC,CAAC,IAAI,EAAE,CAAC;YACnE,MAAM,YAAY,GAAG,GAAG,UAAU,IAAI,SAAS,EAAE,CAAC;YAElD,IAAI,mBAAmB,CAAC,YAAY,CAAC,EAAE,CAAC;gBACtC,SAAS,CAAC,2BAA2B;YACvC,CAAC;YAED,oBAAoB;YACpB,MAAM,OAAO,CAAC,KAAK,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;YAEvC,sBAAsB;YACtB,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;YAE/B,mCAAmC;YACnC,MAAM,cAAc,GAAG;gBACrB,iBAAiB;gBACjB,eAAe;gBACf,QAAQ;gBACR,gBAAgB;gBAChB,WAAW;gBACX,uBAAuB;aACxB,CAAC;YAEF,IAAI,eAAe,GAAG,KAAK,CAAC;YAE5B,KAAK,MAAM,QAAQ,IAAI,cAAc,EAAE,CAAC;gBACtC,MAAM,aAAa,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,GAAG,EAAE,CAAC;gBAEzD,KAAK,MAAM,KAAK,IAAI,aAAa,EAAE,CAAC;oBAClC,IAAI,CAAC;wBACH,MAAM,SAAS,GAAG,MAAM,KAAK,CAAC,SAAS,EAAE,CAAC;wBAC1C,IAAI,SAAS,EAAE,CAAC;4BACd,eAAe,GAAG,IAAI,CAAC;4BAEvB,0CAA0C;4BAC1C,MAAM,aAAa,GAAG,MAAM,gBAAgB,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;4BAE5D,mDAAmD;4BACnD,KAAK,MAAM,IAAI,IAAI,aAAa,CAAC,KAAK,EAAE,CAAC;gCACvC,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC;oCACjE,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gCAC/B,CAAC;4BACH,CAAC;4BAED,KAAK,MAAM,MAAM,IAAI,aAAa,CAAC,OAAO,EAAE,CAAC;gCAC3C,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;oCACrE,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;gCACnC,CAAC;4BACH,CAAC;4BAED,KAAK,MAAM,IAAI,IAAI,aAAa,CAAC,KAAK,EAAE,CAAC;gCACvC,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;oCACzD,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gCAC/B,CAAC;4BACH,CAAC;4BAED,KAAK,MAAM,IAAI,IAAI,aAAa,CAAC,KAAK,EAAE,CAAC;gCACvC,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC;oCACjE,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gCAC/B,CAAC;4BACH,CAAC;4BAED,KAAK,MAAM,KAAK,IAAI,aAAa,CAAC,gBAAgB,EAAE,CAAC;gCACnD,IAAI,CAAC,WAAW,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC;oCAC7E,WAAW,CAAC,gBAAgB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gCAC3C,CAAC;4BACH,CAAC;wBACH,CAAC;oBACH,CAAC;oBAAC,MAAM,CAAC;wBACP,OAAO;oBACT,CAAC;gBACH,CAAC;YACH,CAAC;YAED,qBAAqB;YACrB,IAAI,eAAe,EAAE,CAAC;gBACpB,IAAI,CAAC;oBACH,2CAA2C;oBAC3C,MAAM,OAAO,CAAC,KAAK,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;oBACvC,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;gBACjC,CAAC;gBAAC,MAAM,CAAC;oBACP,sBAAsB;oBACtB,IAAI,CAAC;wBACH,MAAM,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;wBACpC,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;oBACjC,CAAC;oBAAC,MAAM,CAAC;wBACP,iDAAiD;wBACjD,IAAI,CAAC;4BACH,MAAM,IAAI,CAAC,MAAM,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,kBAAkB,EAAE,CAAC,CAAC;wBACtE,CAAC;wBAAC,MAAM,CAAC;4BACP,2BAA2B;4BAC3B,MAAM;wBACR,CAAC;oBACH,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,4CAA4C;YAC5C,SAAS;QACX,CAAC;IACH,CAAC;IAED,uDAAuD;IACvD,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,8BAA8B,CAAC,CAAC,GAAG,EAAE,CAAC;IAE1E,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;QAC5B,IAAI,CAAC;YACH,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,SAAS,EAAE,CAAC;YACzC,IAAI,CAAC,SAAS;gBAAE,SAAS;YAEzB,oBAAoB;YACpB,MAAM,IAAI,CAAC,KAAK,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC;YACpC,MAAM,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC;YAE/B,2BAA2B;YAC3B,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,+BAA+B,CAAC,CAAC,GAAG,EAAE,CAAC;YAE5E,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;gBACjC,IAAI,CAAC;oBACH,MAAM,iBAAiB,GAAG,MAAM,QAAQ,CAAC,SAAS,EAAE,CAAC;oBACrD,IAAI,iBAAiB,EAAE,CAAC;wBACtB,gCAAgC;wBAChC,MAAM,gBAAgB,GAAG,MAAM,gBAAgB,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;wBAE/D,kBAAkB;wBAClB,KAAK,MAAM,IAAI,IAAI,gBAAgB,CAAC,KAAK,EAAE,CAAC;4BAC1C,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gCACzD,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;4BAC/B,CAAC;wBACH,CAAC;wBAED,oBAAoB;wBACpB,KAAK,MAAM,MAAM,IAAI,gBAAgB,CAAC,OAAO,EAAE,CAAC;4BAC9C,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;gCACrE,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;4BACnC,CAAC;wBACH,CAAC;oBACH,CAAC;gBACH,CAAC;gBAAC,MAAM,CAAC;oBACP,OAAO;gBACT,CAAC;YACH,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,4BAA4B;YAC5B,SAAS;QACX,CAAC;IACH,CAAC;IAED,+EAA+E;IAC/E,8FAA8F;IAC9F,qEAAqE;IAErE,OAAO,WAAW,CAAC;AACrB,CAAC"}

package/dist/discovery/index.d.ts

@@ -0,0 +1,8 @@
+export { SiteCrawler } from './crawler.js';
+export type { CrawlerOptions } from './crawler.js';
+export { detectSPAFramework, interceptRouteChanges } from './spa-detector.js';
+export { discoverElements, discoverHiddenElements } from './element-mapper.js';
+export { validateLinks, createLinkValidationState } from './link-validator.js';
+export { buildSitemap, printSitemapTree } from './sitemap-builder.js';
+export { runDiscovery } from './discovery-pipeline.js';
+export type { DiscoveryOptions } from './discovery-pipeline.js';

package/dist/discovery/index.js

@@ -0,0 +1,8 @@
+// Barrel export for discovery module
+export { SiteCrawler } from './crawler.js';
+export { detectSPAFramework, interceptRouteChanges } from './spa-detector.js';
+export { discoverElements, discoverHiddenElements } from './element-mapper.js';
+export { validateLinks, createLinkValidationState } from './link-validator.js';
+export { buildSitemap, printSitemapTree } from './sitemap-builder.js';
+export { runDiscovery } from './discovery-pipeline.js';
+//# sourceMappingURL=index.js.map

package/dist/discovery/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/discovery/index.ts"],"names":[],"mappings":"AAAA,qCAAqC;AACrC,OAAO,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAE3C,OAAO,EAAE,kBAAkB,EAAE,qBAAqB,EAAE,MAAM,mBAAmB,CAAC;AAC9E,OAAO,EAAE,gBAAgB,EAAE,sBAAsB,EAAE,MAAM,qBAAqB,CAAC;AAC/E,OAAO,EAAE,aAAa,EAAE,yBAAyB,EAAE,MAAM,qBAAqB,CAAC;AAC/E,OAAO,EAAE,YAAY,EAAE,gBAAgB,EAAE,MAAM,sBAAsB,CAAC;AACtE,OAAO,EAAE,YAAY,EAAE,MAAM,yBAAyB,CAAC"}

package/dist/discovery/link-validator.d.ts

@@ -0,0 +1,15 @@
+import type { Page } from 'playwright';
+import type { BrokenLink, LinkInfo } from '../types/discovery.js';
+export interface LinkValidationState {
+    checkedCount: number;
+}
+export declare function createLinkValidationState(): LinkValidationState;
+/**
+ * Validates internal links by checking their HTTP status codes.
+ * External links are skipped to avoid rate-limiting third-party sites.
+ *
+ * @param links - All links discovered on a page
+ * @param page - Playwright Page instance (maintains session context for auth)
+ * @returns Array of broken links (4xx, 5xx, network errors)
+ */
+export declare function validateLinks(links: LinkInfo[], page: Page, state?: LinkValidationState): Promise<BrokenLink[]>;

package/dist/discovery/link-validator.js

@@ -0,0 +1,137 @@
+import { ensurePublicHostname } from '../core/validation.js';
+// Security caps for link validation
+const MAX_LINKS_PER_PAGE = 50;
+const MAX_TOTAL_LINKS = 500;
+export function createLinkValidationState() {
+    return { checkedCount: 0 };
+}
+/**
+ * Validates internal links by checking their HTTP status codes.
+ * External links are skipped to avoid rate-limiting third-party sites.
+ *
+ * @param links - All links discovered on a page
+ * @param page - Playwright Page instance (maintains session context for auth)
+ * @returns Array of broken links (4xx, 5xx, network errors)
+ */
+export async function validateLinks(links, page, state = createLinkValidationState()) {
+    const brokenLinks = [];
+    const sourceUrl = page.url();
+    const hostnameSafetyCache = new Map();
+    const assertHostnameIsPublic = async (hostname) => {
+        const normalized = hostname.toLowerCase();
+        if (hostnameSafetyCache.has(normalized)) {
+            const cachedError = hostnameSafetyCache.get(normalized);
+            if (cachedError) {
+                throw new Error(cachedError);
+            }
+            return;
+        }
+        try {
+            await ensurePublicHostname(normalized);
+            hostnameSafetyCache.set(normalized, null);
+        }
+        catch (error) {
+            const message = error instanceof Error ? error.message : String(error);
+            hostnameSafetyCache.set(normalized, message);
+            throw error;
+        }
+    };
+    // Filter to internal links only
+    const internalLinks = links.filter(link => link.isInternal);
+    // Deduplicate by href (normalize query strings)
+    const uniqueUrls = new Map();
+    for (const link of internalLinks) {
+        try {
+            const url = new URL(link.href);
+            // Normalize: remove trailing slash, lowercase hostname
+            const normalizedHref = `${url.origin}${url.pathname.replace(/\/$/, '')}${url.search}`;
+            if (!uniqueUrls.has(normalizedHref)) {
+                uniqueUrls.set(normalizedHref, { ...link, href: normalizedHref });
+            }
+        }
+        catch {
+            // Invalid URL, skip
+            continue;
+        }
+    }
+    // Filter out non-HTTP links
+    const httpLinks = Array.from(uniqueUrls.values()).filter(link => {
+        const href = link.href.toLowerCase();
+        return !href.startsWith('mailto:') &&
+            !href.startsWith('tel:') &&
+            !href.startsWith('javascript:') &&
+            !href.startsWith('data:') &&
+            !href.startsWith('#');
+    });
+    // Apply per-page cap
+    const cappedLinks = httpLinks.slice(0, MAX_LINKS_PER_PAGE);
+    if (httpLinks.length > MAX_LINKS_PER_PAGE) {
+        console.warn(`Capping link validation to ${MAX_LINKS_PER_PAGE} links per page (${httpLinks.length} found)`);
+    }
+    // Apply global cap
+    const remainingGlobalCapacity = MAX_TOTAL_LINKS - state.checkedCount;
+    const linksToCheck = cappedLinks.slice(0, remainingGlobalCapacity);
+    if (linksToCheck.length < cappedLinks.length) {
+        console.warn(`Global link cap reached (${MAX_TOTAL_LINKS} total). Skipping ${cappedLinks.length - linksToCheck.length} links.`);
+    }
+    state.checkedCount += linksToCheck.length;
+    // Check links with concurrency limit (max 10 concurrent)
+    const batchSize = 10;
+    for (let i = 0; i < linksToCheck.length; i += batchSize) {
+        const batch = linksToCheck.slice(i, i + batchSize);
+        const results = await Promise.allSettled(batch.map(async (link) => {
+            try {
+                const initialParsedUrl = new URL(link.href);
+                await assertHostnameIsPublic(initialParsedUrl.hostname);
+                // Use page.request to maintain browser session context (cookies, auth)
+                const response = await page.request.get(link.href, {
+                    timeout: 5000,
+                    // Follow redirects by default - only final status matters
+                });
+                // SSRF protection: check final redirect destination
+                const finalUrl = response.url();
+                if (finalUrl !== link.href) {
+                    try {
+                        const finalParsed = new URL(finalUrl);
+                        await assertHostnameIsPublic(finalParsed.hostname);
+                    }
+                    catch (error) {
+                        return {
+                            url: link.href,
+                            sourceUrl,
+                            statusCode: 0,
+                            statusText: error instanceof Error ? error.message : 'SSRF protection: Redirect destination failed validation',
+                        };
+                    }
+                }
+                // Check if response is broken (4xx, 5xx)
+                if (response.status() >= 400) {
+                    return {
+                        url: link.href,
+                        sourceUrl,
+                        statusCode: response.status(),
+                        statusText: response.statusText(),
+                    };
+                }
+                return null; // Link is OK
+            }
+            catch (error) {
+                // Network error or timeout
+                return {
+                    url: link.href,
+                    sourceUrl,
+                    statusCode: 0,
+                    statusText: error instanceof Error ? error.message : 'Network error',
+                };
+            }
+        }));
+        // Collect broken links from this batch
+        for (const result of results) {
+            if (result.status === 'fulfilled' && result.value !== null) {
+                brokenLinks.push(result.value);
+            }
+        }
+    }
+    return brokenLinks;
+}
+//# sourceMappingURL=link-validator.js.map

package/dist/discovery/link-validator.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"link-validator.js","sourceRoot":"","sources":["../../src/discovery/link-validator.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,oBAAoB,EAAE,MAAM,uBAAuB,CAAC;AAE7D,oCAAoC;AACpC,MAAM,kBAAkB,GAAG,EAAE,CAAC;AAC9B,MAAM,eAAe,GAAG,GAAG,CAAC;AAM5B,MAAM,UAAU,yBAAyB;IACvC,OAAO,EAAE,YAAY,EAAE,CAAC,EAAE,CAAC;AAC7B,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,aAAa,CACjC,KAAiB,EACjB,IAAU,EACV,QAA6B,yBAAyB,EAAE;IAExD,MAAM,WAAW,GAAiB,EAAE,CAAC;IACrC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAC7B,MAAM,mBAAmB,GAAG,IAAI,GAAG,EAAyB,CAAC;IAE7D,MAAM,sBAAsB,GAAG,KAAK,EAAE,QAAgB,EAAiB,EAAE;QACvE,MAAM,UAAU,GAAG,QAAQ,CAAC,WAAW,EAAE,CAAC;QAC1C,IAAI,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,EAAE,CAAC;YACxC,MAAM,WAAW,GAAG,mBAAmB,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;YACxD,IAAI,WAAW,EAAE,CAAC;gBAChB,MAAM,IAAI,KAAK,CAAC,WAAW,CAAC,CAAC;YAC/B,CAAC;YACD,OAAO;QACT,CAAC;QAED,IAAI,CAAC;YACH,MAAM,oBAAoB,CAAC,UAAU,CAAC,CAAC;YACvC,mBAAmB,CAAC,GAAG,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;QAC5C,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,OAAO,GAAG,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;YACvE,mBAAmB,CAAC,GAAG,CAAC,UAAU,EAAE,OAAO,CAAC,CAAC;YAC7C,MAAM,KAAK,CAAC;QACd,CAAC;IACH,CAAC,CAAC;IAEF,gCAAgC;IAChC,MAAM,aAAa,GAAG,KAAK,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IAE5D,gDAAgD;IAChD,MAAM,UAAU,GAAG,IAAI,GAAG,EAAoB,CAAC;IAC/C,KAAK,MAAM,IAAI,IAAI,aAAa,EAAE,CAAC;QACjC,IAAI,CAAC;YACH,MAAM,GAAG,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC/B,uDAAuD;YACvD,MAAM,cAAc,GAAG,GAAG,GAAG,CAAC,MAAM,GAAG,GAAG,CAAC,QAAQ,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;YAEtF,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,cAAc,CAAC,EAAE,CAAC;gBACpC,UAAU,CAAC,GAAG,CAAC,cAAc,EAAE,EAAE,GAAG,IAAI,EAAE,IAAI,EAAE,cAAc,EAAE,CAAC,CAAC;YACpE,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,oBAAoB;YACpB,SAAS;QACX,CAAC;IACH,CAAC;IAED,4BAA4B;IAC5B,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,EAAE;QAC9D,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;QACrC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC;YAC3B,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;YACxB,CAAC,IAAI,CAAC,UAAU,CAAC,aAAa,CAAC;YAC/B,CAAC,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;YACzB,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC;IAC/B,CAAC,CAAC,CAAC;IAEH,qBAAqB;IACrB,MAAM,WAAW,GAAG,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,kBAAkB,CAAC,CAAC;IAC3D,IAAI,SAAS,CAAC,MAAM,GAAG,kBAAkB,EAAE,CAAC;QAC1C,OAAO,CAAC,IAAI,CAAC,8BAA8B,kBAAkB,oBAAoB,SAAS,CAAC,MAAM,SAAS,CAAC,CAAC;IAC9G,CAAC;IAED,mBAAmB;IACnB,MAAM,uBAAuB,GAAG,eAAe,GAAG,KAAK,CAAC,YAAY,CAAC;IACrE,MAAM,YAAY,GAAG,WAAW,CAAC,KAAK,CAAC,CAAC,EAAE,uBAAuB,CAAC,CAAC;IAEnE,IAAI,YAAY,CAAC,MAAM,GAAG,WAAW,CAAC,MAAM,EAAE,CAAC;QAC7C,OAAO,CAAC,IAAI,CAAC,4BAA4B,eAAe,qBAAqB,WAAW,CAAC,MAAM,GAAG,YAAY,CAAC,MAAM,SAAS,CAAC,CAAC;IAClI,CAAC;IAED,KAAK,CAAC,YAAY,IAAI,YAAY,CAAC,MAAM,CAAC;IAE1C,yDAAyD;IACzD,MAAM,SAAS,GAAG,EAAE,CAAC;IACrB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;QACxD,MAAM,KAAK,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC;QAEnD,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,UAAU,CACtC,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE;YACvB,IAAI,CAAC;gBACH,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBAC5C,MAAM,sBAAsB,CAAC,gBAAgB,CAAC,QAAQ,CAAC,CAAC;gBAExD,uEAAuE;gBACvE,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE;oBACjD,OAAO,EAAE,IAAI;oBACb,0DAA0D;iBAC3D,CAAC,CAAC;gBAEH,oDAAoD;gBACpD,MAAM,QAAQ,GAAG,QAAQ,CAAC,GAAG,EAAE,CAAC;gBAChC,IAAI,QAAQ,KAAK,IAAI,CAAC,IAAI,EAAE,CAAC;oBAC3B,IAAI,CAAC;wBACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,CAAC;wBACtC,MAAM,sBAAsB,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAC;oBACrD,CAAC;oBAAC,OAAO,KAAK,EAAE,CAAC;wBACf,OAAO;4BACL,GAAG,EAAE,IAAI,CAAC,IAAI;4BACd,SAAS;4BACT,UAAU,EAAE,CAAC;4BACb,UAAU,EAAE,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,yDAAyD;yBAC/G,CAAC;oBACJ,CAAC;gBACH,CAAC;gBAED,yCAAyC;gBACzC,IAAI,QAAQ,CAAC,MAAM,EAAE,IAAI,GAAG,EAAE,CAAC;oBAC7B,OAAO;wBACL,GAAG,EAAE,IAAI,CAAC,IAAI;wBACd,SAAS;wBACT,UAAU,EAAE,QAAQ,CAAC,MAAM,EAAE;wBAC7B,UAAU,EAAE,QAAQ,CAAC,UAAU,EAAE;qBAClC,CAAC;gBACJ,CAAC;gBAED,OAAO,IAAI,CAAC,CAAC,aAAa;YAC5B,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,2BAA2B;gBAC3B,OAAO;oBACL,GAAG,EAAE,IAAI,CAAC,IAAI;oBACd,SAAS;oBACT,UAAU,EAAE,CAAC;oBACb,UAAU,EAAE,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,eAAe;iBACrE,CAAC;YACJ,CAAC;QACH,CAAC,CAAC,CACH,CAAC;QAEF,uCAAuC;QACvC,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;YAC7B,IAAI,MAAM,CAAC,MAAM,KAAK,WAAW,IAAI,MAAM,CAAC,KAAK,KAAK,IAAI,EAAE,CAAC;gBAC3D,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;YACjC,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,WAAW,CAAC;AACrB,CAAC"}

package/dist/discovery/sitemap-builder.d.ts

@@ -0,0 +1,19 @@
+import type { SitemapNode, PageData } from '../types/discovery.js';
+/**
+ * Builds a hierarchical sitemap tree from flat page data based on URL path structure.
+ * Creates intermediate placeholder nodes for uncrawled parent paths.
+ *
+ * @param pages - Flat array of crawled pages
+ * @param rootUrl - Root URL of the site (e.g., https://example.com)
+ * @returns Root node of the sitemap tree
+ */
+export declare function buildSitemap(pages: PageData[], rootUrl: string): SitemapNode;
+/**
+ * Renders a sitemap tree as human-readable indented text.
+ * Format: Title (path) with 2 spaces per depth level.
+ *
+ * @param node - Root or any node in the tree
+ * @param indent - Current indentation level (internal use)
+ * @returns Multi-line string representation of the tree
+ */
+export declare function printSitemapTree(node: SitemapNode, indent?: number): string;