@ni-c/imap-mcp 0.2.0 → 0.4.0

package/dist/config.js

@@ -23,6 +23,18 @@ const DEFAULT_MAX_MESSAGES = 100;
 const DEFAULT_MAX_ATTACHMENT_BYTES = 1024 * 1024;
 /** Disk cap: this one protects the filesystem, which is a different concern. */
 const DEFAULT_MAX_DOWNLOAD_BYTES = 25 * 1024 * 1024;
+/** Extraction cap: how much hostile input one parser is asked to chew on. */
+const DEFAULT_MAX_EXTRACT_BYTES = 10 * 1024 * 1024;
+/**
+ * Hard ceiling on `IMAP_MAX_EXTRACT_BYTES`.
+ *
+ * The other two size variables have no maximum because their cost is linear and
+ * paid somewhere visible — a big result, a big file. This one buys a buffer
+ * inside a parser working on bytes a stranger chose, so a typo of "104857600000"
+ * would not produce a big answer, it would produce an operator who believes
+ * there is a limit. 64 MiB is past any document that arrives by mail.
+ */
+const MAX_MAX_EXTRACT_BYTES = 64 * 1024 * 1024;
 const DEFAULT_SEEN_KEYWORD = 'AiSeen';
 /** Shown when the configuration is incomplete — at startup and on every call. */
 export function missingConfigMessage(missing) {
@@ -60,6 +72,10 @@ export function loadConfig(env = process.env) {
     // to child processes and in /proc/<pid>/environ.
     delete env.IMAP_PASSWORD;
     const tls = parseTlsMode(env.IMAP_TLS);
+    // After the password delete, deliberately: this one can exit the process, and
+    // an exit above the delete would leave the password in the environment for
+    // whatever runs next.
+    const elicitation = parseElicitation(env.ELICITATION);
     if (host !== undefined)
         assertSafeHost(host, 'IMAP_HOST');
     const draftsMailbox = env.IMAP_DRAFTS_MAILBOX;
@@ -87,9 +103,11 @@ export function loadConfig(env = process.env) {
             allowedAttachmentTypes: parseTypes(env.IMAP_ATTACHMENT_TYPES),
             downloadDir: env.IMAP_DOWNLOAD_DIR,
             maxDownloadBytes: parseCount(env.IMAP_MAX_DOWNLOAD_BYTES, DEFAULT_MAX_DOWNLOAD_BYTES, 'IMAP_MAX_DOWNLOAD_BYTES'),
+            maxExtractBytes: parseCount(env.IMAP_MAX_EXTRACT_BYTES, DEFAULT_MAX_EXTRACT_BYTES, 'IMAP_MAX_EXTRACT_BYTES', MAX_MAX_EXTRACT_BYTES),
         },
         // Defaults to true, unlike the rest of the family — see the field comment.
         readOnly: env.IMAP_READ_ONLY !== 'false',
+        elicitation,
         allowTools: env.IMAP_ALLOW_TOOLS,
         denyTools: env.IMAP_DENY_TOOLS,
     };
@@ -122,6 +140,29 @@ function parseTlsMode(raw) {
     console.error('imap-mcp: IMAP_TLS must be one of implicit, starttls or none');
     process.exit(1);
 }
+/**
+ * Reads `ELICITATION` — deliberately unprefixed, and deliberately fatal on
+ * anything it does not recognise.
+ *
+ * Unprefixed: environment variables are process-wide, so this is one switch for
+ * every server in the same environment. That is also its risk, which is why a
+ * server started with it off says so on its startup line.
+ *
+ * Fatal, like `parseTlsMode` above and unlike `IMAP_READ_ONLY`: this is the
+ * first variable of the family that defaults to *on*, so a typo that fell back
+ * to the default would leave the dialog running while the operator believes it
+ * is off — and an operator who believes that has no way to find out.
+ */
+export function parseElicitation(raw) {
+    const value = raw?.trim().toLowerCase();
+    if (value === undefined || value === '' || value === 'true')
+        return true;
+    if (value === 'false')
+        return false;
+    console.error(`imap-mcp: ELICITATION must be "true" or "false" — got "${raw}". ` +
+        'Refusing to start rather than guess.');
+    process.exit(1);
+}
 function parsePort(raw, fallback, name) {
     if (raw === undefined || raw === '')
         return fallback;
@@ -133,7 +174,7 @@ function parsePort(raw, fallback, name) {
     }
     return value;
 }
-function parseCount(raw, fallback, name) {
+function parseCount(raw, fallback, name, max) {
     if (raw === undefined || raw === '')
         return fallback;
     const value = Number(raw);
@@ -141,6 +182,13 @@ function parseCount(raw, fallback, name) {
         console.error(`imap-mcp: ${name} must be a positive integer`);
         process.exit(1);
     }
+    if (max !== undefined && value > max) {
+        // The limit is named rather than clamped to: an operator who asked for more
+        // than the server will do should learn that here, not from a refusal later
+        // that looks like the document was the problem.
+        console.error(`imap-mcp: ${name} must not exceed ${max}`);
+        process.exit(1);
+    }
     return value;
 }
 /**
@@ -191,12 +239,23 @@ function assertSingleLine(value, name) {
     }
 }
 function isLoopbackHost(hostname) {
+    // URL.hostname keeps the brackets around an IPv6 literal, may carry a %zone
+    // suffix, and 'localhost.' with its root label is the same name as
+    // 'localhost'. The comparison this replaced saw none of them — which is why
+    // its bare '::1' branch could never match a hostname taken from a URL.
     if (hostname === undefined)
         return false;
-    return (hostname === 'localhost' ||
-        hostname.endsWith('.localhost') ||
-        hostname.startsWith('127.') ||
-        hostname === '::1' ||
-        hostname === '[::1]');
+    const host = hostname
+        .toLowerCase()
+        .replace(/^\[|]$/g, '')
+        .replace(/%.*$/, '')
+        .replace(/\.+$/, '');
+    return (host === 'localhost' ||
+        host.endsWith('.localhost') ||
+        host.startsWith('127.') ||
+        host === '::1' ||
+        // Every dual-stack client dials ::ffff:127.0.0.1 as plain 127.0.0.1, and
+        // URL normalises the mapped form to hex (::ffff:7f00:1).
+        /^::ffff:(?:7f[0-9a-f]{0,2}:|127\.)/.test(host));
 }
 //# sourceMappingURL=config.js.map

package/dist/config.js.map

@@ -1 +1 @@
-{"version":3,"file":"config.js","sourceRoot":"","sources":["../src/config.ts"],"names":[],"mappings":"AA+DA,MAAM,CAAC,MAAM,wBAAwB,GAAG;IACtC,iBAAiB;IACjB,kBAAkB;IAClB,iBAAiB;IACjB,iBAAiB;IACjB,iBAAiB;IACjB,yCAAyC;IACzC,gDAAgD;IAChD,yEAAyE;IACzE,mEAAmE;IACnE,2EAA2E;IAC3E,WAAW;IACX,YAAY;IACZ,WAAW;IACX,YAAY;IACZ,YAAY;IACZ,UAAU;IACV,WAAW;IACX,eAAe;CAChB,CAAC;AAEF,MAAM,oBAAoB,GAAG,GAAG,CAAC;AACjC,gEAAgE;AAChE,MAAM,4BAA4B,GAAG,IAAI,GAAG,IAAI,CAAC;AACjD,gFAAgF;AAChF,MAAM,0BAA0B,GAAG,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;AACpD,MAAM,oBAAoB,GAAG,QAAQ,CAAC;AAEtC,iFAAiF;AACjF,MAAM,UAAU,oBAAoB,CAAC,OAAiB;IACpD,OAAO,CACL,6CAA6C,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI;QACnE,yEAAyE;QACzE,wEAAwE;QACxE,sEAAsE;QACtE,4EAA4E;QAC5E,wEAAwE;QACxE,2DAA2D,CAC5D,CAAC;AACJ,CAAC;AAED,8EAA8E;AAC9E,MAAM,UAAU,iBAAiB,CAAC,MAAc;IAC9C,OAAO;QACL,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,IAAI,WAAW;QAChC,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,IAAI,WAAW;QAChC,CAAC,MAAM,CAAC,IAAI,CAAC,QAAQ,IAAI,eAAe;KACzC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAe,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC;AAC3C,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,UAAU,CAAC,MAAyB,OAAO,CAAC,GAAG;IAC7D,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC;IAC3B,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC;IAC3B,MAAM,QAAQ,GAAG,GAAG,CAAC,aAAa,CAAC;IAEnC,8EAA8E;IAC9E,4EAA4E;IAC5E,iDAAiD;IACjD,OAAO,GAAG,CAAC,aAAa,CAAC;IAEzB,MAAM,GAAG,GAAG,YAAY,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IAEvC,IAAI,IAAI,KAAK,SAAS;QAAE,cAAc,CAAC,IAAI,EAAE,WAAW,CAAC,CAAC;IAC1D,MAAM,aAAa,GAAG,GAAG,CAAC,mBAAmB,CAAC;IAC9C,IAAI,aAAa,KAAK,SAAS,EAAE,CAAC;QAChC,gBAAgB,CAAC,aAAa,EAAE,qBAAqB,CAAC,CAAC;IACzD,CAAC;IACD,MAAM,iBAAiB,GAAG,GAAG,CAAC,wBAAwB,EAAE,IAAI,EAAE,IAAI,SAAS,CAAC;IAC5E,IAAI,iBAAiB,KAAK,SAAS,EAAE,CAAC;QACpC,gBAAgB,CAAC,iBAAiB,EAAE,0BAA0B,CAAC,CAAC;IAClE,CAAC;IAED,MAAM,MAAM,GAAW;QACrB,IAAI,EAAE;YACJ,IAAI;YACJ,IAAI,EAAE,SAAS,CACb,GAAG,CAAC,SAAS,EACb,GAAG,KAAK,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAC9B,WAAW,CACZ;YACD,IAAI;YACJ,QAAQ;YACR,GAAG;YACH,WAAW,EAAE,GAAG,CAAC,iBAAiB,KAAK,MAAM;YAC7C,OAAO,EAAE,GAAG,CAAC,YAAY,IAAI,OAAO;YACpC,WAAW,EAAE,YAAY,CAAC,GAAG,CAAC,iBAAiB,CAAC;YAChD,aAAa;YACb,iBAAiB;YACjB,WAAW,EAAE,UAAU,CACrB,GAAG,CAAC,iBAAiB,EACrB,oBAAoB,EACpB,mBAAmB,CACpB;YACD,kBAAkB,EAAE,UAAU,CAC5B,GAAG,CAAC,yBAAyB,EAC7B,4BAA4B,EAC5B,2BAA2B,CAC5B;YACD,sBAAsB,EAAE,UAAU,CAAC,GAAG,CAAC,qBAAqB,CAAC;YAC7D,WAAW,EAAE,GAAG,CAAC,iBAAiB;YAClC,gBAAgB,EAAE,UAAU,CAC1B,GAAG,CAAC,uBAAuB,EAC3B,0BAA0B,EAC1B,yBAAyB,CAC1B;SACF;QACD,2EAA2E;QAC3E,QAAQ,EAAE,GAAG,CAAC,cAAc,KAAK,OAAO;QACxC,UAAU,EAAE,GAAG,CAAC,gBAAgB;QAChC,SAAS,EAAE,GAAG,CAAC,eAAe;KAC/B,CAAC;IAEF,6EAA6E;IAC7E,2EAA2E;IAC3E,4EAA4E;IAC5E,6EAA6E;IAC7E,iCAAiC;IACjC,IAAI,GAAG,CAAC,gBAAgB,KAAK,SAAS,EAAE,CAAC;QACvC,OAAO,CAAC,KAAK,CACX,sEAAsE;YACpE,sEAAsE;YACtE,gCAAgC,CACnC,CAAC;QACF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IAED,MAAM,OAAO,GAAG,iBAAiB,CAAC,MAAM,CAAC,CAAC;IAC1C,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACvB,OAAO,CAAC,KAAK,CAAC,aAAa,oBAAoB,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IAC9D,CAAC;IACD,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,KAAK,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,EAAE,CAAC;QACxD,OAAO,CAAC,KAAK,CACX,2EAA2E;YACzE,uDAAuD,CAC1D,CAAC;IACJ,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,YAAY,CAAC,GAAuB;IAC3C,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,UAAU,CAAC;IACvD,IAAI,GAAG,KAAK,UAAU,IAAI,GAAG,KAAK,UAAU,IAAI,GAAG,KAAK,MAAM;QAAE,OAAO,GAAG,CAAC;IAC3E,OAAO,CAAC,KAAK,CAAC,8DAA8D,CAAC,CAAC;IAC9E,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC;AAED,SAAS,SAAS,CAChB,GAAuB,EACvB,QAAgB,EAChB,IAAY;IAEZ,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,QAAQ,CAAC;IACrD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;IAC1B,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,IAAI,KAAK,GAAG,KAAK,EAAE,CAAC;QAC3D,gEAAgE;QAChE,OAAO,CAAC,KAAK,CAAC,aAAa,IAAI,yCAAyC,CAAC,CAAC;QAC1E,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,UAAU,CACjB,GAAuB,EACvB,QAAgB,EAChB,IAAY;IAEZ,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,QAAQ,CAAC;IACrD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;IAC1B,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;QAC1C,OAAO,CAAC,KAAK,CAAC,aAAa,IAAI,6BAA6B,CAAC,CAAC;QAC9D,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;GAGG;AACH,SAAS,YAAY,CAAC,GAAuB;IAC3C,IAAI,GAAG,KAAK,SAAS;QAAE,OAAO,oBAAoB,CAAC;IACnD,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,EAAE,CAAC;IAC1B,IAAI,CAAC,oBAAoB,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QACpC,OAAO,CAAC,KAAK,CACX,2EAA2E,CAC5E,CAAC;QACF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,UAAU,CAAC,GAAuB;IACzC,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,CAAC,IAAI,EAAE,KAAK,EAAE;QAAE,OAAO,wBAAwB,CAAC;IAC5E,OAAO,GAAG;SACP,KAAK,CAAC,GAAG,CAAC;SACV,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;SAClC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC;AAC7B,CAAC;AAED;;;;GAIG;AACH,SAAS,cAAc,CAAC,KAAa,EAAE,IAAY;IACjD,2EAA2E;IAC3E,kEAAkE;IAClE,sEAAsE;IACtE,MAAM,QAAQ,GAAG,mBAAmB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACjD,MAAM,IAAI,GAAG,uCAAuC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACjE,IAAI,CAAC,QAAQ,IAAI,CAAC,IAAI,EAAE,CAAC;QACvB,OAAO,CAAC,KAAK,CACX,aAAa,IAAI,kDAAkD;YACjE,yCAAyC,CAC5C,CAAC;QACF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;AACH,CAAC;AAED,6EAA6E;AAC7E,SAAS,gBAAgB,CAAC,KAAa,EAAE,IAAY;IACnD,IAAI,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,CAAC,KAAK,CAAC,aAAa,IAAI,+BAA+B,CAAC,CAAC;QAChE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;AACH,CAAC;AAED,SAAS,cAAc,CAAC,QAA4B;IAClD,IAAI,QAAQ,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACzC,OAAO,CACL,QAAQ,KAAK,WAAW;QACxB,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC;QAC/B,QAAQ,CAAC,UAAU,CAAC,MAAM,CAAC;QAC3B,QAAQ,KAAK,KAAK;QAClB,QAAQ,KAAK,OAAO,CACrB,CAAC;AACJ,CAAC"}
+{"version":3,"file":"config.js","sourceRoot":"","sources":["../src/config.ts"],"names":[],"mappings":"AAkFA,MAAM,CAAC,MAAM,wBAAwB,GAAG;IACtC,iBAAiB;IACjB,kBAAkB;IAClB,iBAAiB;IACjB,iBAAiB;IACjB,iBAAiB;IACjB,yCAAyC;IACzC,gDAAgD;IAChD,yEAAyE;IACzE,mEAAmE;IACnE,2EAA2E;IAC3E,WAAW;IACX,YAAY;IACZ,WAAW;IACX,YAAY;IACZ,YAAY;IACZ,UAAU;IACV,WAAW;IACX,eAAe;CAChB,CAAC;AAEF,MAAM,oBAAoB,GAAG,GAAG,CAAC;AACjC,gEAAgE;AAChE,MAAM,4BAA4B,GAAG,IAAI,GAAG,IAAI,CAAC;AACjD,gFAAgF;AAChF,MAAM,0BAA0B,GAAG,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;AACpD,6EAA6E;AAC7E,MAAM,yBAAyB,GAAG,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;AACnD;;;;;;;;GAQG;AACH,MAAM,qBAAqB,GAAG,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;AAC/C,MAAM,oBAAoB,GAAG,QAAQ,CAAC;AAEtC,iFAAiF;AACjF,MAAM,UAAU,oBAAoB,CAAC,OAAiB;IACpD,OAAO,CACL,6CAA6C,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI;QACnE,yEAAyE;QACzE,wEAAwE;QACxE,sEAAsE;QACtE,4EAA4E;QAC5E,wEAAwE;QACxE,2DAA2D,CAC5D,CAAC;AACJ,CAAC;AAED,8EAA8E;AAC9E,MAAM,UAAU,iBAAiB,CAAC,MAAc;IAC9C,OAAO;QACL,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,IAAI,WAAW;QAChC,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,IAAI,WAAW;QAChC,CAAC,MAAM,CAAC,IAAI,CAAC,QAAQ,IAAI,eAAe;KACzC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAe,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC;AAC3C,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,UAAU,CAAC,GAAG,GAAsB,OAAO,CAAC,GAAG;IAC7D,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC;IAC3B,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC;IAC3B,MAAM,QAAQ,GAAG,GAAG,CAAC,aAAa,CAAC;IAEnC,8EAA8E;IAC9E,4EAA4E;IAC5E,iDAAiD;IACjD,OAAO,GAAG,CAAC,aAAa,CAAC;IAEzB,MAAM,GAAG,GAAG,YAAY,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;IACvC,8EAA8E;IAC9E,2EAA2E;IAC3E,sBAAsB;IACtB,MAAM,WAAW,GAAG,gBAAgB,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;IAEtD,IAAI,IAAI,KAAK,SAAS;QAAE,cAAc,CAAC,IAAI,EAAE,WAAW,CAAC,CAAC;IAC1D,MAAM,aAAa,GAAG,GAAG,CAAC,mBAAmB,CAAC;IAC9C,IAAI,aAAa,KAAK,SAAS,EAAE,CAAC;QAChC,gBAAgB,CAAC,aAAa,EAAE,qBAAqB,CAAC,CAAC;IACzD,CAAC;IACD,MAAM,iBAAiB,GAAG,GAAG,CAAC,wBAAwB,EAAE,IAAI,EAAE,IAAI,SAAS,CAAC;IAC5E,IAAI,iBAAiB,KAAK,SAAS,EAAE,CAAC;QACpC,gBAAgB,CAAC,iBAAiB,EAAE,0BAA0B,CAAC,CAAC;IAClE,CAAC;IAED,MAAM,MAAM,GAAW;QACrB,IAAI,EAAE;YACJ,IAAI;YACJ,IAAI,EAAE,SAAS,CACb,GAAG,CAAC,SAAS,EACb,GAAG,KAAK,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAC9B,WAAW,CACZ;YACD,IAAI;YACJ,QAAQ;YACR,GAAG;YACH,WAAW,EAAE,GAAG,CAAC,iBAAiB,KAAK,MAAM;YAC7C,OAAO,EAAE,GAAG,CAAC,YAAY,IAAI,OAAO;YACpC,WAAW,EAAE,YAAY,CAAC,GAAG,CAAC,iBAAiB,CAAC;YAChD,aAAa;YACb,iBAAiB;YACjB,WAAW,EAAE,UAAU,CACrB,GAAG,CAAC,iBAAiB,EACrB,oBAAoB,EACpB,mBAAmB,CACpB;YACD,kBAAkB,EAAE,UAAU,CAC5B,GAAG,CAAC,yBAAyB,EAC7B,4BAA4B,EAC5B,2BAA2B,CAC5B;YACD,sBAAsB,EAAE,UAAU,CAAC,GAAG,CAAC,qBAAqB,CAAC;YAC7D,WAAW,EAAE,GAAG,CAAC,iBAAiB;YAClC,gBAAgB,EAAE,UAAU,CAC1B,GAAG,CAAC,uBAAuB,EAC3B,0BAA0B,EAC1B,yBAAyB,CAC1B;YACD,eAAe,EAAE,UAAU,CACzB,GAAG,CAAC,sBAAsB,EAC1B,yBAAyB,EACzB,wBAAwB,EACxB,qBAAqB,CACtB;SACF;QACD,2EAA2E;QAC3E,QAAQ,EAAE,GAAG,CAAC,cAAc,KAAK,OAAO;QACxC,WAAW;QACX,UAAU,EAAE,GAAG,CAAC,gBAAgB;QAChC,SAAS,EAAE,GAAG,CAAC,eAAe;KAC/B,CAAC;IAEF,6EAA6E;IAC7E,2EAA2E;IAC3E,4EAA4E;IAC5E,6EAA6E;IAC7E,iCAAiC;IACjC,IAAI,GAAG,CAAC,gBAAgB,KAAK,SAAS,EAAE,CAAC;QACvC,OAAO,CAAC,KAAK,CACX,sEAAsE;YACpE,sEAAsE;YACtE,gCAAgC,CACnC,CAAC;QACF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IAED,MAAM,OAAO,GAAG,iBAAiB,CAAC,MAAM,CAAC,CAAC;IAC1C,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACvB,OAAO,CAAC,KAAK,CAAC,aAAa,oBAAoB,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IAC9D,CAAC;IACD,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,KAAK,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,EAAE,CAAC;QACxD,OAAO,CAAC,KAAK,CACX,2EAA2E;YACzE,uDAAuD,CAC1D,CAAC;IACJ,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,YAAY,CAAC,GAAuB;IAC3C,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,UAAU,CAAC;IACvD,IAAI,GAAG,KAAK,UAAU,IAAI,GAAG,KAAK,UAAU,IAAI,GAAG,KAAK,MAAM;QAAE,OAAO,GAAG,CAAC;IAC3E,OAAO,CAAC,KAAK,CAAC,8DAA8D,CAAC,CAAC;IAC9E,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC;AAED;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,gBAAgB,CAAC,GAAuB;IACtD,MAAM,KAAK,GAAG,GAAG,EAAE,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IACxC,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,EAAE,IAAI,KAAK,KAAK,MAAM;QAAE,OAAO,IAAI,CAAC;IACzE,IAAI,KAAK,KAAK,OAAO;QAAE,OAAO,KAAK,CAAC;IACpC,OAAO,CAAC,KAAK,CACX,0DAA0D,GAAG,KAAK;QAChE,sCAAsC,CACzC,CAAC;IACF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC;AAED,SAAS,SAAS,CAChB,GAAuB,EACvB,QAAgB,EAChB,IAAY;IAEZ,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,QAAQ,CAAC;IACrD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;IAC1B,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,IAAI,KAAK,GAAG,KAAK,EAAE,CAAC;QAC3D,gEAAgE;QAChE,OAAO,CAAC,KAAK,CAAC,aAAa,IAAI,yCAAyC,CAAC,CAAC;QAC1E,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,UAAU,CACjB,GAAuB,EACvB,QAAgB,EAChB,IAAY,EACZ,GAAY;IAEZ,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,QAAQ,CAAC;IACrD,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;IAC1B,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;QAC1C,OAAO,CAAC,KAAK,CAAC,aAAa,IAAI,6BAA6B,CAAC,CAAC;QAC9D,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IACD,IAAI,GAAG,KAAK,SAAS,IAAI,KAAK,GAAG,GAAG,EAAE,CAAC;QACrC,4EAA4E;QAC5E,2EAA2E;QAC3E,gDAAgD;QAChD,OAAO,CAAC,KAAK,CAAC,aAAa,IAAI,oBAAoB,GAAG,EAAE,CAAC,CAAC;QAC1D,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;GAGG;AACH,SAAS,YAAY,CAAC,GAAuB;IAC3C,IAAI,GAAG,KAAK,SAAS;QAAE,OAAO,oBAAoB,CAAC;IACnD,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,EAAE,CAAC;IAC1B,IAAI,CAAC,oBAAoB,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;QACpC,OAAO,CAAC,KAAK,CACX,2EAA2E,CAC5E,CAAC;QACF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,UAAU,CAAC,GAAuB;IACzC,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,CAAC,IAAI,EAAE,KAAK,EAAE;QAAE,OAAO,wBAAwB,CAAC;IAC5E,OAAO,GAAG;SACP,KAAK,CAAC,GAAG,CAAC;SACV,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;SAClC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC;AAC7B,CAAC;AAED;;;;GAIG;AACH,SAAS,cAAc,CAAC,KAAa,EAAE,IAAY;IACjD,2EAA2E;IAC3E,kEAAkE;IAClE,sEAAsE;IACtE,MAAM,QAAQ,GAAG,mBAAmB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACjD,MAAM,IAAI,GAAG,uCAAuC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACjE,IAAI,CAAC,QAAQ,IAAI,CAAC,IAAI,EAAE,CAAC;QACvB,OAAO,CAAC,KAAK,CACX,aAAa,IAAI,kDAAkD;YACjE,yCAAyC,CAC5C,CAAC;QACF,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;AACH,CAAC;AAED,6EAA6E;AAC7E,SAAS,gBAAgB,CAAC,KAAa,EAAE,IAAY;IACnD,IAAI,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,CAAC,KAAK,CAAC,aAAa,IAAI,+BAA+B,CAAC,CAAC;QAChE,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC;AACH,CAAC;AAED,SAAS,cAAc,CAAC,QAA4B;IAClD,4EAA4E;IAC5E,mEAAmE;IACnE,4EAA4E;IAC5E,uEAAuE;IACvE,IAAI,QAAQ,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACzC,MAAM,IAAI,GAAG,QAAQ;SAClB,WAAW,EAAE;SACb,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC;SACtB,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC;SACnB,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IACvB,OAAO,CACL,IAAI,KAAK,WAAW;QACpB,IAAI,CAAC,QAAQ,CAAC,YAAY,CAAC;QAC3B,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QACvB,IAAI,KAAK,KAAK;QACd,yEAAyE;QACzE,yDAAyD;QACzD,oCAAoC,CAAC,IAAI,CAAC,IAAI,CAAC,CAChD,CAAC;AACJ,CAAC"}

package/dist/errors.js.map

@@ -1 +1 @@
-{"version":3,"file":"errors.js","sourceRoot":"","sources":["../src/errors.ts"],"names":[],"mappings":"AAAA,gFAAgF;AAChF,MAAM,OAAO,cAAe,SAAQ,KAAK;IACvC,YAAY,OAAe;QACzB,KAAK,CAAC,OAAO,CAAC,CAAC;QACf,IAAI,CAAC,IAAI,GAAG,gBAAgB,CAAC;IAC/B,CAAC;CACF;AAED;;;;;GAKG;AACH,MAAM,OAAO,SAAU,SAAQ,KAAK;IAGvB;IACA;IAHX,YACE,OAAe,EACN,OAA2B,SAAS,EACpC,eAAuB,EAAE;QAElC,KAAK,CAAC,OAAO,CAAC,CAAC;QAHN,SAAI,GAAJ,IAAI,CAAgC;QACpC,iBAAY,GAAZ,YAAY,CAAa;QAGlC,IAAI,CAAC,IAAI,GAAG,WAAW,CAAC;IAC1B,CAAC;CACF"}
+{"version":3,"file":"errors.js","sourceRoot":"","sources":["../src/errors.ts"],"names":[],"mappings":"AAAA,gFAAgF;AAChF,MAAM,OAAO,cAAe,SAAQ,KAAK;IACvC,YAAY,OAAe;QACzB,KAAK,CAAC,OAAO,CAAC,CAAC;QACf,IAAI,CAAC,IAAI,GAAG,gBAAgB,CAAC;IAC/B,CAAC;CACF;AAED;;;;;GAKG;AACH,MAAM,OAAO,SAAU,SAAQ,KAAK;IAGvB,IAAI;IACJ,YAAY;IAHvB,YACE,OAAe,EACN,IAAI,GAAuB,SAAS,EACpC,YAAY,GAAW,EAAE;QAElC,KAAK,CAAC,OAAO,CAAC,CAAC;oBAHN,IAAI;4BACJ,YAAY;QAGrB,IAAI,CAAC,IAAI,GAAG,WAAW,CAAC;IAC1B,CAAC;CACF"}

package/dist/extract/child.d.ts

@@ -0,0 +1 @@
+export {};

package/dist/extract/child.js

@@ -0,0 +1,83 @@
+/**
+ * The extraction child: the only code in this server that parses a binary a
+ * stranger sent, and the reason it runs in a process of its own.
+ *
+ * A process rather than a worker thread, and that was a lesson rather than a
+ * preference. A worker's `resourceLimits` promise to turn a runaway parse into
+ * `ERR_WORKER_OUT_OF_MEMORY`, and for most allocation patterns they do; for
+ * the one a 3 kB spreadsheet produced they did not, and the whole server died
+ * with `FATAL ERROR: Reached heap limit`. A thread also cannot give back the
+ * memory a terminated parse left behind — a gigabyte stayed resident after
+ * `terminate()`. A process that is killed takes its memory with it, whatever
+ * it was doing, and a process that aborts aborts alone.
+ *
+ * Read the import rules before editing this file. It is started from its own
+ * source — `child.ts` under vitest, which Node runs with type stripping, and
+ * `child.js` from `dist/` in production — so Node loads it and everything it
+ * reaches, without vitest's resolver. Node strips types; it does not rewrite a
+ * `./x.js` specifier to the `./x.ts` sitting beside it. So:
+ *
+ * - static imports here may only be bare specifiers (`node:*`, a package) or
+ *   `import type`, which is erased before it can fail to resolve;
+ * - anything relative is loaded through {@link sibling}, which picks the
+ *   extension from this module's own;
+ * - no enums, no namespaces, no parameter properties — type stripping cannot
+ *   erase syntax that emits code.
+ *
+ * Get this wrong in one direction and every test fails while the build stays
+ * green; get it wrong in the other and every test passes while the published
+ * package throws on first use. The second is why `npm run build` is followed by
+ * a real extraction against `dist/`.
+ */
+const EXTENSION = import.meta.url.endsWith('.ts') ? 'ts' : 'js';
+function sibling(path) {
+    return new URL(`${path}.${EXTENSION}`, import.meta.url).href;
+}
+async function run(request) {
+    if (request.kind === 'pdf') {
+        const { extractPdf } = (await import(sibling('./pdf')));
+        return extractPdf(request.bytes, request.maxChars);
+    }
+    const [{ extractZipDocument }, { htmlToText }] = await Promise.all([
+        import(sibling('./ooxml')),
+        import(sibling('../analyze')),
+    ]);
+    return extractZipDocument(request.kind, request.bytes, request.maxChars, htmlToText);
+}
+/** A code for the log, never a message: the message quotes the document. */
+function codeOf(error) {
+    const value = error;
+    if (typeof value?.code === 'string')
+        return value.code;
+    if (typeof value?.name === 'string')
+        return value.name;
+    return 'unknown';
+}
+const send = process.send?.bind(process);
+if (send !== undefined) {
+    // The parent is gone: nothing is waiting for an answer, and a parse that
+    // outlives the server it was started by is a parse nobody asked for.
+    process.on('disconnect', () => {
+        process.exit(0);
+    });
+    process.once('message', (request) => {
+        run(request).then((response) => {
+            send(response, () => {
+                process.exit(0);
+            });
+        }, (error) => {
+            // Never the caught error. pdf.js and fflate quote the document in
+            // their exception messages — byte offsets, object fragments, what they
+            // found where they expected something else. That is text a stranger
+            // wrote, and an error message is read as the server's own voice,
+            // outside the fence every other piece of message content passes
+            // through. A code crosses; the words are the host's.
+            console.error(`imap-mcp: extraction failed: ${codeOf(error)}`);
+            send({ ok: false, reason: 'internal' }, () => {
+                process.exit(0);
+            });
+        });
+    });
+}
+export {};
+//# sourceMappingURL=child.js.map

package/dist/extract/child.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"child.js","sourceRoot":"","sources":["../../src/extract/child.ts"],"names":[],"mappings":"AAEA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA8BG;AACH,MAAM,SAAS,GAAG,OAAO,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;AAEhE,SAAS,OAAO,CAAC,IAAY;IAC3B,OAAO,IAAI,GAAG,CAAC,GAAG,IAAI,IAAI,SAAS,EAAE,EAAE,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC;AAC/D,CAAC;AAED,KAAK,UAAU,GAAG,CAAC,OAAuB;IACxC,IAAI,OAAO,CAAC,IAAI,KAAK,KAAK,EAAE,CAAC;QAC3B,MAAM,EAAE,UAAU,EAAE,GAAG,CAAC,MAAM,MAAM,CAClC,OAAO,CAAC,OAAO,CAAC,CACjB,CAA8B,CAAC;QAChC,OAAO,UAAU,CAAC,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,QAAQ,CAAC,CAAC;IACrD,CAAC;IACD,MAAM,CAAC,EAAE,kBAAkB,EAAE,EAAE,EAAE,UAAU,EAAE,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QACjE,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,CAAyC;QAClE,MAAM,CAAC,OAAO,CAAC,YAAY,CAAC,CAA4C;KACzE,CAAC,CAAC;IACH,OAAO,kBAAkB,CACvB,OAAO,CAAC,IAAI,EACZ,OAAO,CAAC,KAAK,EACb,OAAO,CAAC,QAAQ,EAChB,UAAU,CACX,CAAC;AACJ,CAAC;AAED,4EAA4E;AAC5E,SAAS,MAAM,CAAC,KAAc;IAC5B,MAAM,KAAK,GAAG,KAAkD,CAAC;IACjE,IAAI,OAAO,KAAK,EAAE,IAAI,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC,IAAI,CAAC;IACvD,IAAI,OAAO,KAAK,EAAE,IAAI,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC,IAAI,CAAC;IACvD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;AACzC,IAAI,IAAI,KAAK,SAAS,EAAE,CAAC;IACvB,yEAAyE;IACzE,qEAAqE;IACrE,OAAO,CAAC,EAAE,CAAC,YAAY,EAAE,GAAG,EAAE;QAC5B,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAClB,CAAC,CAAC,CAAC;IACH,OAAO,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC,OAAuB,EAAE,EAAE;QAClD,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CACf,CAAC,QAAQ,EAAE,EAAE;YACX,IAAI,CAAC,QAAQ,EAAE,GAAG,EAAE;gBAClB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAClB,CAAC,CAAC,CAAC;QACL,CAAC,EACD,CAAC,KAAc,EAAE,EAAE;YACjB,kEAAkE;YAClE,uEAAuE;YACvE,oEAAoE;YACpE,iEAAiE;YACjE,gEAAgE;YAChE,qDAAqD;YACrD,OAAO,CAAC,KAAK,CAAC,gCAAgC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;YAC/D,IAAI,CACF,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,UAAU,EAA4B,EAC3D,GAAG,EAAE;gBACH,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YAClB,CAAC,CACF,CAAC;QACJ,CAAC,CACF,CAAC;IACJ,CAAC,CAAC,CAAC;AACL,CAAC"}

package/dist/extract/index.d.ts

@@ -0,0 +1,41 @@
+import type { ExtractKind, ExtractRequest, ExtractResponse } from './types.js';
+export type { ExtractKind, ExtractReason, ExtractRequest, ExtractResponse, } from './types.js';
+/**
+ * How long a document may be parsed before the process doing it is killed.
+ *
+ * Generous next to the five seconds a regular expression gets in the sibling
+ * servers, because a hundred-page PDF is honest work — and finite, because a
+ * document that has not finished by now is not going to.
+ */
+export declare const EXTRACT_TIMEOUT_MS = 20000;
+/**
+ * Characters the child may return, before any of the paging below.
+ *
+ * Not a context budget — that is `max_chars` at the tool, and it is much
+ * smaller. This is the ceiling on what is held in memory and paged through,
+ * and the child enforces it for every format: nothing larger is ever built
+ * there, let alone sent back.
+ */
+export declare const MAX_EXTRACT_CHARS = 1000000;
+/** The same set as content types, for `get_server_info`. */
+export declare const EXTRACTABLE_TYPES: string[];
+/** Prose for the refusals, so every one of them names the same set. */
+export declare const EXTRACTABLE_TYPE_NAMES: string;
+export declare function extractKindOf(contentType: string): ExtractKind | undefined;
+export declare function isExtractable(contentType: string): boolean;
+/** The magic-byte verdict an honest container of this kind produces. */
+export declare function expectedSignature(kind: ExtractKind): string;
+export declare function extractDocumentText(request: ExtractRequest, limits?: ChildLimits): Promise<ExtractResponse>;
+/**
+ * Narrowed by the tests, never by the server.
+ *
+ * The timeout and the memory ceiling are the two guards whose whole purpose is
+ * what happens when they fire, and neither can be reached in a test at its real
+ * value without a document engineered to spend twenty seconds or a quarter of a
+ * gigabyte. Making them arguments is what lets the failure paths be exercised
+ * in milliseconds; nothing in `src/` passes them.
+ */
+export interface ChildLimits {
+    timeoutMs?: number;
+    memoryMb?: number;
+}

package/dist/extract/index.js

@@ -0,0 +1,183 @@
+import { fork } from 'node:child_process';
+import { once } from 'node:events';
+/**
+ * How long a document may be parsed before the process doing it is killed.
+ *
+ * Generous next to the five seconds a regular expression gets in the sibling
+ * servers, because a hundred-page PDF is honest work — and finite, because a
+ * document that has not finished by now is not going to.
+ */
+export const EXTRACT_TIMEOUT_MS = 20_000;
+/**
+ * Characters the child may return, before any of the paging below.
+ *
+ * Not a context budget — that is `max_chars` at the tool, and it is much
+ * smaller. This is the ceiling on what is held in memory and paged through,
+ * and the child enforces it for every format: nothing larger is ever built
+ * there, let alone sent back.
+ */
+export const MAX_EXTRACT_CHARS = 1_000_000;
+/**
+ * V8 heap the parsing process may use.
+ *
+ * Best effort, and named as such. It bounds a pathological object graph, and
+ * when it fires the child aborts — on its own, which is the whole reason the
+ * parse is in a process. It does not bound typed arrays, which are external
+ * memory; the deflate pre-scan in `pdf.ts` and the entry caps in `ooxml.ts`
+ * are what cover those.
+ */
+const CHILD_MEMORY_MB = 256;
+/**
+ * Requests admitted at once, running and waiting together.
+ *
+ * Extractions run one at a time (see {@link queue}), so a request that arrives
+ * behind seven others would wait up to seven timeouts for its turn, holding
+ * its mailbox lock throughout. Past this many it is refused outright, which is
+ * an answer the caller can act on.
+ */
+const MAX_IN_FLIGHT = 8;
+/** Content types this server can read text out of, and what each one is. */
+const EXTRACTABLE = new Map([
+    ['application/pdf', 'pdf'],
+    [
+        'application/vnd.openxmlformats-officedocument.wordprocessingml.document',
+        'docx',
+    ],
+    ['application/vnd.openxmlformats-officedocument.spreadsheetml.sheet', 'xlsx'],
+    [
+        'application/vnd.openxmlformats-officedocument.presentationml.presentation',
+        'pptx',
+    ],
+    ['application/vnd.oasis.opendocument.text', 'odt'],
+    ['application/vnd.oasis.opendocument.spreadsheet', 'ods'],
+]);
+/** The same set as content types, for `get_server_info`. */
+export const EXTRACTABLE_TYPES = [...EXTRACTABLE.keys()];
+/** Prose for the refusals, so every one of them names the same set. */
+export const EXTRACTABLE_TYPE_NAMES = 'PDF, Word (.docx), Excel (.xlsx), PowerPoint (.pptx) and OpenDocument text ' +
+    'and spreadsheets (.odt, .ods)';
+export function extractKindOf(contentType) {
+    return EXTRACTABLE.get(contentType.toLowerCase());
+}
+export function isExtractable(contentType) {
+    return EXTRACTABLE.has(contentType.toLowerCase());
+}
+/** The magic-byte verdict an honest container of this kind produces. */
+export function expectedSignature(kind) {
+    // Every OOXML and OpenDocument file is a zip, which is why `sniffContent`
+    // reports one for all five of them.
+    return kind === 'pdf' ? 'application/pdf' : 'application/zip';
+}
+/**
+ * Serialises extractions.
+ *
+ * One process per call and no limit would mean N concurrent tool calls holding
+ * N processes of {@link CHILD_MEMORY_MB} each, which is a memory limit that
+ * multiplies by a number the caller chooses. The queue is the whole mechanism:
+ * the next extraction starts when the previous process is gone.
+ */
+let queue = Promise.resolve();
+let inFlight = 0;
+export async function extractDocumentText(request, limits = {}) {
+    if (inFlight >= MAX_IN_FLIGHT)
+        return { ok: false, reason: 'busy' };
+    inFlight += 1;
+    try {
+        const run = queue.then(() => runInChild(request, limits), () => runInChild(request, limits));
+        queue = run.catch(() => undefined);
+        return await run;
+    }
+    finally {
+        inFlight -= 1;
+    }
+}
+/** A code for the log, never a message. */
+function codeOf(error) {
+    const value = error;
+    if (typeof value?.code === 'string')
+        return value.code;
+    if (typeof value?.name === 'string')
+        return value.name;
+    return 'unknown';
+}
+async function runInChild(request, limits = {}) {
+    const timeoutMs = limits.timeoutMs ?? EXTRACT_TIMEOUT_MS;
+    const memoryMb = limits.memoryMb ?? CHILD_MEMORY_MB;
+    const child = fork(new URL(import.meta.url.endsWith('.ts') ? './child.ts' : './child.js', import.meta.url), [], {
+        // Stated rather than inherited. The parent's own flags may be ones a
+        // child cannot take — `--input-type` is one — and an inherited flag that
+        // fails to parse would turn every extraction into a silent `internal`.
+        execArgv: [`--max-old-space-size=${memoryMb}`],
+        // Not tidiness — correctness. The parent's stdout is this server's
+        // JSON-RPC transport, and pdf.js logs. One line from inside the parser
+        // would corrupt the framing and hang the session. Discarded; stderr is
+        // shared, because that is where every other diagnostic in this server
+        // already goes.
+        stdio: ['ignore', 'ignore', 'inherit', 'ipc'],
+        // Structured clone rather than JSON: the request carries the document as
+        // a typed array, and JSON would turn it into an array of numbers ten
+        // times its size.
+        serialization: 'advanced',
+    });
+    let timer;
+    try {
+        return await new Promise((resolve) => {
+            // The first answer wins. Everything after it — the exit of a child that
+            // was killed because it had already answered, most of all — is silence,
+            // not a second verdict.
+            let settled = false;
+            const settle = (value) => {
+                if (settled)
+                    return;
+                settled = true;
+                resolve(value);
+            };
+            timer = setTimeout(() => {
+                settle({ ok: false, reason: 'timeout' });
+            }, timeoutMs);
+            child.once('message', (value) => {
+                settle(value);
+            });
+            child.once('error', (error) => {
+                if (settled)
+                    return;
+                console.error(`imap-mcp: extraction process failed: ${codeOf(error)}`);
+                settle({ ok: false, reason: 'internal' });
+            });
+            child.once('exit', (code, signal) => {
+                // Only reached when the child left without answering. An abort is what
+                // V8 does when the heap limit is hit — and what the process does
+                // *instead of* taking the server with it, which is the property being
+                // bought here.
+                if (settled)
+                    return;
+                if (signal === 'SIGABRT' || code === 134) {
+                    settle({ ok: false, reason: 'out-of-memory' });
+                    return;
+                }
+                console.error(`imap-mcp: extraction process exited early: ${signal ?? `code ${code}`}`);
+                settle({ ok: false, reason: 'internal' });
+            });
+            try {
+                child.send(request);
+            }
+            catch (error) {
+                console.error(`imap-mcp: extraction request failed: ${codeOf(error)}`);
+                settle({ ok: false, reason: 'internal' });
+            }
+        });
+    }
+    finally {
+        if (timer)
+            clearTimeout(timer);
+        // Unconditional: on the timeout path the process is still inside the parse
+        // and will never exit on its own. SIGKILL, because a parser stuck in native
+        // code does not check for anything gentler — and because a process, unlike
+        // a thread, can be killed from outside whatever it is doing.
+        if (child.exitCode === null && child.signalCode === null) {
+            child.kill('SIGKILL');
+            await once(child, 'exit');
+        }
+    }
+}
+//# sourceMappingURL=index.js.map

package/dist/extract/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/extract/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,IAAI,EAAE,MAAM,oBAAoB,CAAC;AAC1C,OAAO,EAAE,IAAI,EAAE,MAAM,aAAa,CAAC;AAWnC;;;;;;GAMG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAG,MAAM,CAAC;AAEzC;;;;;;;GAOG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG,SAAS,CAAC;AAE3C;;;;;;;;GAQG;AACH,MAAM,eAAe,GAAG,GAAG,CAAC;AAE5B;;;;;;;GAOG;AACH,MAAM,aAAa,GAAG,CAAC,CAAC;AAExB,4EAA4E;AAC5E,MAAM,WAAW,GAAG,IAAI,GAAG,CAAsB;IAC/C,CAAC,iBAAiB,EAAE,KAAK,CAAC;IAC1B;QACE,yEAAyE;QACzE,MAAM;KACP;IACD,CAAC,mEAAmE,EAAE,MAAM,CAAC;IAC7E;QACE,2EAA2E;QAC3E,MAAM;KACP;IACD,CAAC,yCAAyC,EAAE,KAAK,CAAC;IAClD,CAAC,gDAAgD,EAAE,KAAK,CAAC;CAC1D,CAAC,CAAC;AAEH,4DAA4D;AAC5D,MAAM,CAAC,MAAM,iBAAiB,GAAa,CAAC,GAAG,WAAW,CAAC,IAAI,EAAE,CAAC,CAAC;AAEnE,uEAAuE;AACvE,MAAM,CAAC,MAAM,sBAAsB,GACjC,6EAA6E;IAC7E,+BAA+B,CAAC;AAElC,MAAM,UAAU,aAAa,CAAC,WAAmB;IAC/C,OAAO,WAAW,CAAC,GAAG,CAAC,WAAW,CAAC,WAAW,EAAE,CAAC,CAAC;AACpD,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,WAAmB;IAC/C,OAAO,WAAW,CAAC,GAAG,CAAC,WAAW,CAAC,WAAW,EAAE,CAAC,CAAC;AACpD,CAAC;AAED,wEAAwE;AACxE,MAAM,UAAU,iBAAiB,CAAC,IAAiB;IACjD,0EAA0E;IAC1E,oCAAoC;IACpC,OAAO,IAAI,KAAK,KAAK,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,iBAAiB,CAAC;AAChE,CAAC;AAED;;;;;;;GAOG;AACH,IAAI,KAAK,GAAqB,OAAO,CAAC,OAAO,EAAE,CAAC;AAChD,IAAI,QAAQ,GAAG,CAAC,CAAC;AAEjB,MAAM,CAAC,KAAK,UAAU,mBAAmB,CACvC,OAAuB,EACvB,MAAM,GAAgB,EAAE;IAExB,IAAI,QAAQ,IAAI,aAAa;QAAE,OAAO,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC;IACpE,QAAQ,IAAI,CAAC,CAAC;IACd,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CACpB,GAAG,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,MAAM,CAAC,EACjC,GAAG,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,MAAM,CAAC,CAClC,CAAC;QACF,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;QACnC,OAAO,MAAM,GAAG,CAAC;IACnB,CAAC;YAAS,CAAC;QACT,QAAQ,IAAI,CAAC,CAAC;IAChB,CAAC;AACH,CAAC;AAgBD,2CAA2C;AAC3C,SAAS,MAAM,CAAC,KAAc;IAC5B,MAAM,KAAK,GAAG,KAAkD,CAAC;IACjE,IAAI,OAAO,KAAK,EAAE,IAAI,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC,IAAI,CAAC;IACvD,IAAI,OAAO,KAAK,EAAE,IAAI,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC,IAAI,CAAC;IACvD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,KAAK,UAAU,UAAU,CACvB,OAAuB,EACvB,MAAM,GAAgB,EAAE;IAExB,MAAM,SAAS,GAAG,MAAM,CAAC,SAAS,IAAI,kBAAkB,CAAC;IACzD,MAAM,QAAQ,GAAG,MAAM,CAAC,QAAQ,IAAI,eAAe,CAAC;IAEpD,MAAM,KAAK,GAAG,IAAI,CAChB,IAAI,GAAG,CACL,OAAO,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,YAAY,EAC7D,OAAO,IAAI,CAAC,GAAG,CAChB,EACD,EAAE,EACF;QACE,qEAAqE;QACrE,yEAAyE;QACzE,uEAAuE;QACvE,QAAQ,EAAE,CAAC,wBAAwB,QAAQ,EAAE,CAAC;QAC9C,mEAAmE;QACnE,uEAAuE;QACvE,uEAAuE;QACvE,sEAAsE;QACtE,gBAAgB;QAChB,KAAK,EAAE,CAAC,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,KAAK,CAAC;QAC7C,yEAAyE;QACzE,qEAAqE;QACrE,kBAAkB;QAClB,aAAa,EAAE,UAAU;KAC1B,CACF,CAAC;IAEF,IAAI,KAAiC,CAAC;IACtC,IAAI,CAAC;QACH,OAAO,MAAM,IAAI,OAAO,CAAkB,CAAC,OAAO,EAAE,EAAE;YACpD,wEAAwE;YACxE,wEAAwE;YACxE,wBAAwB;YACxB,IAAI,OAAO,GAAG,KAAK,CAAC;YACpB,MAAM,MAAM,GAAG,CAAC,KAAsB,EAAQ,EAAE;gBAC9C,IAAI,OAAO;oBAAE,OAAO;gBACpB,OAAO,GAAG,IAAI,CAAC;gBACf,OAAO,CAAC,KAAK,CAAC,CAAC;YACjB,CAAC,CAAC;YACF,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE;gBACtB,MAAM,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,SAAS,EAAE,CAAC,CAAC;YAC3C,CAAC,EAAE,SAAS,CAAC,CAAC;YACd,KAAK,CAAC,IAAI,CAAC,SAAS,EAAE,CAAC,KAAK,EAAE,EAAE;gBAC9B,MAAM,CAAC,KAAwB,CAAC,CAAC;YACnC,CAAC,CAAC,CAAC;YACH,KAAK,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC,KAAK,EAAE,EAAE;gBAC5B,IAAI,OAAO;oBAAE,OAAO;gBACpB,OAAO,CAAC,KAAK,CAAC,wCAAwC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;gBACvE,MAAM,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,UAAU,EAAE,CAAC,CAAC;YAC5C,CAAC,CAAC,CAAC;YACH,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,EAAE,MAAM,EAAE,EAAE;gBAClC,uEAAuE;gBACvE,iEAAiE;gBACjE,sEAAsE;gBACtE,eAAe;gBACf,IAAI,OAAO;oBAAE,OAAO;gBACpB,IAAI,MAAM,KAAK,SAAS,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;oBACzC,MAAM,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,eAAe,EAAE,CAAC,CAAC;oBAC/C,OAAO;gBACT,CAAC;gBACD,OAAO,CAAC,KAAK,CACX,8CAA8C,MAAM,IAAI,QAAQ,IAAI,EAAE,EAAE,CACzE,CAAC;gBACF,MAAM,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,UAAU,EAAE,CAAC,CAAC;YAC5C,CAAC,CAAC,CAAC;YACH,IAAI,CAAC;gBACH,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YACtB,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,OAAO,CAAC,KAAK,CAAC,wCAAwC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;gBACvE,MAAM,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,UAAU,EAAE,CAAC,CAAC;YAC5C,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;YAAS,CAAC;QACT,IAAI,KAAK;YAAE,YAAY,CAAC,KAAK,CAAC,CAAC;QAC/B,2EAA2E;QAC3E,4EAA4E;QAC5E,2EAA2E;QAC3E,6DAA6D;QAC7D,IAAI,KAAK,CAAC,QAAQ,KAAK,IAAI,IAAI,KAAK,CAAC,UAAU,KAAK,IAAI,EAAE,CAAC;YACzD,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;YACtB,MAAM,IAAI,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;QAC5B,CAAC;IACH,CAAC;AACH,CAAC"}

package/dist/extract/ooxml.d.ts

@@ -0,0 +1,35 @@
+import type { ExtractKind, ExtractResponse } from './types.js';
+/**
+ * Turns markup into readable text. Injected rather than imported.
+ *
+ * `htmlToText` lives in `../analyze.ts`, and this module is reached from the
+ * extraction child, which Node loads directly — where a `../analyze.js`
+ * specifier does not resolve to the `.ts` beside it. Passing the function in
+ * keeps this file free of relative value imports, which is the property that
+ * lets it run in the child at all. It also makes the seam explicit: the tests
+ * hand it the same `htmlToText` the child does.
+ */
+export type MarkupToText = (markup: string, maxChars: number) => string;
+/**
+ * Reads the text of an OOXML or OpenDocument container.
+ *
+ * The whole defence lives in the `filter` callback below, and it is worth
+ * saying why that specific place. `unzipSync` inflates an entry into a buffer
+ * sized by the *declared* uncompressed size out of the central directory — a
+ * number the sender chose, checked against nothing. The filter is the last
+ * point before that allocation, and returning `false` there means the entry is
+ * never inflated and never sized.
+ *
+ * Measured on fflate 0.8.3: a declared size far past the real one does not blow
+ * up resident memory on Linux, because the allocation is virtual and untouched
+ * pages cost nothing; and a declared size *below* the real one truncates the
+ * output to what was declared, because fflate does not grow a caller-sized
+ * buffer. The guard stays regardless — it is free, it is the only thing
+ * standing between an *honest* high-ratio entry and its real expansion, and a
+ * host that does not overcommit would pay the full price.
+ *
+ * This never recurses. An entry that is itself an archive is not in the name
+ * allowlist, so a nested bomb is not descended into; that is a property to keep
+ * rather than an omission to fix.
+ */
+export declare function extractZipDocument(kind: Exclude<ExtractKind, 'pdf'>, bytes: Uint8Array, maxChars: number, toText: MarkupToText): Promise<ExtractResponse>;