@monoes/monomindcli 2.9.2 → 2.9.3

package/dist/src/utils/redaction.js

@@ -0,0 +1,190 @@
+/**
+ * Shared redaction — paths, PII, and secret patterns.
+ *
+ * Previously duplicated in three places that had already diverged:
+ *   - crash-reporter.ts's `redact()` — the fullest implementation (POSIX +
+ *     Windows paths, IPv4/IPv6, email, SSN, phone, 12 secret patterns,
+ *     hostnames). This module is a lift of that pipeline.
+ *   - input-guards.ts's `sanitizeError()` — a single POSIX-only path regex,
+ *     no PII/secret handling at all.
+ *   - neural-optimize.ts's `stripPii` block — POSIX-only `/Users`/`/home`
+ *     paths, email, IPv4; no Windows paths, no secret patterns, despite
+ *     defaulting ON specifically because it feeds a publicly-published
+ *     pattern export.
+ * Consolidated here so a new secret pattern or a Windows-path fix reaches
+ * every caller instead of two of three.
+ */
+import { homedir } from 'os';
+// Step order matters within `redact()` — see the inline comments below for
+// why (mirrors crash-reporter.ts's original pipeline exactly).
+function stripHomeAndUser(text, home) {
+    let out = text;
+    if (home)
+        out = out.split(home).join('~');
+    const username = home.split('/').pop();
+    if (username && username.length > 2) {
+        out = out.replace(new RegExp(`\\b${username}\\b`, 'g'), '<user>');
+    }
+    return out;
+}
+// Generic path prefixes, in case a compiled binary's stack trace embeds a
+// *different* machine's home dir (e.g. the CI runner or maintainer's build
+// box) than the one stripHomeAndUser() is keyed to. Includes the Windows
+// form — the gap both former narrow implementations had.
+function stripGenericUserPrefixes(text) {
+    let out = text;
+    out = out.replace(/\/home\/[^/\s]+/g, '/home/<user>');
+    out = out.replace(/\/Users\/[^/\s]+/g, '/Users/<user>');
+    out = out.replace(/C:\\Users\\[^\\\s]+/g, 'C:\\Users\\<user>');
+    return out;
+}
+function stripIPv4(text) {
+    return text.replace(/\b(?:\d{1,3}\.){3}\d{1,3}\b/g, '<ip>');
+}
+// #124-review: the original `[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}`
+// has the same catastrophic-backtracking shape as the hostname regex fixed
+// above, in TWO places: the unbounded local-part quantifier backtracks
+// character-by-character at every scan position when no `@` follows (a
+// long dotted string with no `@` at all — ~2s on 40K chars), and the
+// domain part's `[A-Za-z0-9.-]+` vs the final `\.[A-Za-z]{2,}` has the
+// exact middle-vs-final split ambiguity the hostname fix removes. Local
+// part bounded to 64 chars (RFC 5321's actual limit); domain rewritten so
+// each label's trailing dot is consumed by the SAME repeated group, never
+// re-attemptable as part of the final TLD segment, with the repetition
+// itself capped at 10 labels (real hostnames essentially never need more).
+function stripEmail(text) {
+    return text.replace(/\b[A-Za-z0-9._%+-]{1,64}@(?:[A-Za-z0-9](?:[A-Za-z0-9-]{0,61}[A-Za-z0-9])?\.){1,10}[A-Za-z]{2,24}\b/g, '<email>');
+}
+function stripSsnAndPhone(text) {
+    let out = text;
+    out = out.replace(/\b\d{3}-\d{2}-\d{4}\b/g, '<ssn>');
+    out = out.replace(/(?:\+?\d{1,3}[-.\s]?)?\(?\d{3}\)?[-.\s]?\d{3}[-.\s]?\d{4}\b/g, '<phone>');
+    return out;
+}
+const SECRET_PATTERNS = [
+    /(?:api[_-]?key|apikey)\s*[:=]\s*['"]?[^\s'"]{8,}['"]?/gi,
+    /(?:secret|password|passwd|pwd)\s*[:=]\s*['"]?[^\s'"]{8,}['"]?/gi,
+    /(?:token|bearer)\s*[:=]\s*['"]?[^\s'"]{10,}['"]?/gi,
+    /-----BEGIN (?:RSA |EC |DSA |OPENSSH )?PRIVATE KEY-----[\s\S]*?-----END (?:RSA |EC |DSA |OPENSSH )?PRIVATE KEY-----/g,
+    /sk-ant-[a-zA-Z0-9_-]{20,}/g,
+    /sk-[a-zA-Z0-9_-]{20,}/g,
+    /ghp_[a-zA-Z0-9]{36}/g,
+    /gho_[a-zA-Z0-9]{36}/g,
+    /npm_[a-zA-Z0-9]{36}/g,
+    /AKIA[0-9A-Z]{16}/g,
+    /eyJ[a-zA-Z0-9_-]{10,}\.[a-zA-Z0-9_-]{10,}\.[a-zA-Z0-9_-]{10,}/g, // JWT
+    /[a-zA-Z][a-zA-Z0-9+.-]*:\/\/[^:\s]+:[^@\s]+@[^\s'"]+/g, // user:pass@host connection strings
+];
+function stripSecretPatterns(text) {
+    let out = text;
+    for (const pattern of SECRET_PATTERNS)
+        out = out.replace(pattern, '[redacted]');
+    return out;
+}
+// Collapse any remaining multi-segment absolute path to its basename,
+// preserving optional :line:col suffixes for debuggability. Fires on
+// /a/b, ~/a/b, C:\a\b — single-segment paths (/tmp) are not touched.
+// Must run AFTER home-dir/username replacement — see redactPaths()/redact().
+function collapseAbsolutePaths(text) {
+    return text.replace(/(?:~\/|(?<![:/>\w])\/(?!\/)|[A-Z]:\\)(?:[^\s'"]*[/\\])([\w.-]+(?::\d+(?::\d+)?)?)/g, '<path>/$1');
+}
+const KNOWN_CODE_EXTS = /\.(?:js|ts|jsx|tsx|json|md|yaml|yml|html|css|mjs|cjs|toml|xml|txt|log|sql|sh|py|go|rs|rb|java|c|cpp|h|hpp|swift|kt|lock|map)$/i;
+function stripHostnames(text) {
+    let out = text;
+    // Contextual: hostnames that follow :// or @ (requires at least one dot).
+    out = out.replace(/(\/\/|@)([a-zA-Z0-9](?:[a-zA-Z0-9-]*[a-zA-Z0-9])?(?:\.[a-zA-Z0-9](?:[a-zA-Z0-9-]*[a-zA-Z0-9])?)+)/g, '$1<host>');
+    // Standalone: 3+ dot-separated segments with a TLD-like tail (2-6 alpha
+    // chars). Excludes version numbers (v2.9.0, 1.2.3), filenames whose last
+    // segment is a known code/config extension, and bare decimal numbers.
+    //
+    // #124-review: the original two-part regex (`(?:\.seg)+` immediately
+    // followed by a required `\.[a-zA-Z]{2,6}`) let the engine backtrack over
+    // every way to split a long run of dot-segments between "middle
+    // repetitions" and "the final segment" before concluding no match —
+    // catastrophic (measured ~quadratic) backtracking on a crafted non-
+    // matching input (e.g. a long chain of short dotted tokens with no valid
+    // trailing TLD), reachable via both crash-reporter's redact() and
+    // neural-optimize.ts's public pattern-export path. Rewritten so segments
+    // can never contain a dot themselves — there is exactly one way to parse
+    // any input, so there is nothing to backtrack over. The "last segment
+    // must be a short alpha TLD" check moves into the callback instead of the
+    // regex, since the regex itself no longer distinguishes middle vs final.
+    out = out.replace(/\b[a-zA-Z][a-zA-Z0-9-]*(?:\.[a-zA-Z0-9][a-zA-Z0-9-]*){2,}\b/g, (match) => {
+        if (/^v?\d+(\.\d+)+$/.test(match))
+            return match; // version
+        if (KNOWN_CODE_EXTS.test(match))
+            return match; // filename
+        if (/^\d+(\.\d+)+$/.test(match))
+            return match; // decimal
+        const lastSegment = match.slice(match.lastIndexOf('.') + 1);
+        if (!/^[a-zA-Z]{2,6}$/.test(lastSegment))
+            return match; // not a TLD-shaped tail
+        return '<host>';
+    });
+    return out;
+}
+function stripIPv6(text) {
+    let out = text;
+    // Bracketed form (URLs): [::1], [fe80::1], [2001:db8::1]
+    out = out.replace(/\[(?:[0-9a-fA-F]{0,4}:){2,}[0-9a-fA-F]{0,4}\]/g, '[<ipv6>]');
+    // Full unbracketed: 3+ colon-separated hex groups (avoids HH:MM:SS false positives)
+    out = out.replace(/(?<![:\w])(?:[0-9a-fA-F]{1,4}:){3,7}[0-9a-fA-F]{0,4}(?![:\w])/g, '<ipv6>');
+    // Compressed :: with leading hex groups: fe80::1, 2001:db8::1.
+    // Uses \b (not lookbehind) so hex digits before :: don't block the match.
+    // std::vector is safe: 's' is not [0-9a-fA-F].
+    out = out.replace(/\b[0-9a-fA-F]{1,4}(?::[0-9a-fA-F]{1,4})*::(?:[0-9a-fA-F]{1,4}(?::[0-9a-fA-F]{1,4})*)?\b/g, '<ipv6>');
+    // Compressed :: without leading hex: ::1, ::ffff:10.0.0.1
+    out = out.replace(/(?<![:\w])::(?:[0-9a-fA-F]{1,4}(?::[0-9a-fA-F]{1,4})*)?(?![:\w])/g, '<ipv6>');
+    return out;
+}
+/**
+ * Path redaction: home dir, username, generic `/home`/`/Users`/`C:\Users`
+ * prefixes (POSIX and Windows), and a generic absolute-path-to-basename
+ * collapse. Standalone entry point for callers that only need paths gone
+ * (e.g. sanitizeError() — an internal error message, not published data).
+ */
+export function redactPaths(text) {
+    const home = homedir();
+    let out = stripHomeAndUser(text, home);
+    out = stripGenericUserPrefixes(out);
+    out = collapseAbsolutePaths(out);
+    return out;
+}
+/** IPv4/IPv6, email, SSN, phone, and hostname redaction. */
+export function redactPii(text) {
+    let out = stripIPv4(text);
+    out = stripEmail(out);
+    out = stripSsnAndPhone(out);
+    out = stripHostnames(out);
+    out = stripIPv6(out);
+    return out;
+}
+/** API keys, tokens, passwords, private keys, and known provider key formats. */
+export function redactSecrets(text) {
+    return stripSecretPatterns(text);
+}
+/**
+ * Strip obvious secrets/PII/paths before anything gets sent somewhere it
+ * shouldn't (a public GitHub issue, an exported pattern file). Not a
+ * substitute for careful callers — this is a last-resort net.
+ *
+ * Order matters: home/user/generic-prefix replacement must run before the
+ * generic absolute-path collapse (so a username-bearing path collapses to
+ * `<user>`'s neighborhood correctly), and secret-pattern stripping runs
+ * before that path collapse too (a secret string that happens to look
+ * path-like should still hit the secret patterns first).
+ */
+export function redact(text) {
+    const home = homedir();
+    let out = stripHomeAndUser(text, home);
+    out = stripGenericUserPrefixes(out);
+    out = stripIPv4(out);
+    out = stripEmail(out);
+    out = stripSsnAndPhone(out);
+    out = stripSecretPatterns(out);
+    out = collapseAbsolutePaths(out);
+    out = stripHostnames(out);
+    out = stripIPv6(out);
+    return out;
+}
+//# sourceMappingURL=redaction.js.map

package/dist/src/utils/redaction.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"redaction.js","sourceRoot":"","sources":["../../../src/utils/redaction.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AACH,OAAO,EAAE,OAAO,EAAE,MAAM,IAAI,CAAC;AAE7B,2EAA2E;AAC3E,+DAA+D;AAE/D,SAAS,gBAAgB,CAAC,IAAY,EAAE,IAAY;IAClD,IAAI,GAAG,GAAG,IAAI,CAAC;IACf,IAAI,IAAI;QAAE,GAAG,GAAG,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAE1C,MAAM,QAAQ,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC;IACvC,IAAI,QAAQ,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACpC,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,IAAI,MAAM,CAAC,MAAM,QAAQ,KAAK,EAAE,GAAG,CAAC,EAAE,QAAQ,CAAC,CAAC;IACpE,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,0EAA0E;AAC1E,2EAA2E;AAC3E,yEAAyE;AACzE,yDAAyD;AACzD,SAAS,wBAAwB,CAAC,IAAY;IAC5C,IAAI,GAAG,GAAG,IAAI,CAAC;IACf,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,kBAAkB,EAAE,cAAc,CAAC,CAAC;IACtD,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,mBAAmB,EAAE,eAAe,CAAC,CAAC;IACxD,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,sBAAsB,EAAE,mBAAmB,CAAC,CAAC;IAC/D,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,IAAI,CAAC,OAAO,CAAC,8BAA8B,EAAE,MAAM,CAAC,CAAC;AAC9D,CAAC;AAED,6EAA6E;AAC7E,2EAA2E;AAC3E,uEAAuE;AACvE,uEAAuE;AACvE,qEAAqE;AACrE,uEAAuE;AACvE,wEAAwE;AACxE,0EAA0E;AAC1E,0EAA0E;AAC1E,uEAAuE;AACvE,2EAA2E;AAC3E,SAAS,UAAU,CAAC,IAAY;IAC9B,OAAO,IAAI,CAAC,OAAO,CACjB,qGAAqG,EACrG,SAAS,CACV,CAAC;AACJ,CAAC;AAED,SAAS,gBAAgB,CAAC,IAAY;IACpC,IAAI,GAAG,GAAG,IAAI,CAAC;IACf,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,wBAAwB,EAAE,OAAO,CAAC,CAAC;IACrD,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,8DAA8D,EAAE,SAAS,CAAC,CAAC;IAC7F,OAAO,GAAG,CAAC;AACb,CAAC;AAED,MAAM,eAAe,GAAa;IAChC,yDAAyD;IACzD,iEAAiE;IACjE,oDAAoD;IACpD,qHAAqH;IACrH,4BAA4B;IAC5B,wBAAwB;IACxB,sBAAsB;IACtB,sBAAsB;IACtB,sBAAsB;IACtB,mBAAmB;IACnB,gEAAgE,EAAE,MAAM;IACxE,uDAAuD,EAAS,oCAAoC;CACrG,CAAC;AAEF,SAAS,mBAAmB,CAAC,IAAY;IACvC,IAAI,GAAG,GAAG,IAAI,CAAC;IACf,KAAK,MAAM,OAAO,IAAI,eAAe;QAAE,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,YAAY,CAAC,CAAC;IAChF,OAAO,GAAG,CAAC;AACb,CAAC;AAED,sEAAsE;AACtE,qEAAqE;AACrE,qEAAqE;AACrE,6EAA6E;AAC7E,SAAS,qBAAqB,CAAC,IAAY;IACzC,OAAO,IAAI,CAAC,OAAO,CACjB,oFAAoF,EACpF,WAAW,CACZ,CAAC;AACJ,CAAC;AAED,MAAM,eAAe,GAAG,gIAAgI,CAAC;AAEzJ,SAAS,cAAc,CAAC,IAAY;IAClC,IAAI,GAAG,GAAG,IAAI,CAAC;IACf,0EAA0E;IAC1E,GAAG,GAAG,GAAG,CAAC,OAAO,CACf,oGAAoG,EACpG,UAAU,CACX,CAAC;IACF,wEAAwE;IACxE,yEAAyE;IACzE,sEAAsE;IACtE,EAAE;IACF,qEAAqE;IACrE,0EAA0E;IAC1E,gEAAgE;IAChE,oEAAoE;IACpE,oEAAoE;IACpE,yEAAyE;IACzE,kEAAkE;IAClE,yEAAyE;IACzE,yEAAyE;IACzE,sEAAsE;IACtE,0EAA0E;IAC1E,yEAAyE;IACzE,GAAG,GAAG,GAAG,CAAC,OAAO,CACf,8DAA8D,EAC9D,CAAC,KAAK,EAAE,EAAE;QACR,IAAI,iBAAiB,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,OAAO,KAAK,CAAC,CAAO,UAAU;QACjE,IAAI,eAAe,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,OAAO,KAAK,CAAC,CAAU,WAAW;QACnE,IAAI,eAAe,CAAC,IAAI,CAAC,KAAK,CAAC;YAAE,OAAO,KAAK,CAAC,CAAU,UAAU;QAClE,MAAM,WAAW,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;QAC5D,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,WAAW,CAAC;YAAE,OAAO,KAAK,CAAC,CAAC,wBAAwB;QAChF,OAAO,QAAQ,CAAC;IAClB,CAAC,CACF,CAAC;IACF,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,SAAS,CAAC,IAAY;IAC7B,IAAI,GAAG,GAAG,IAAI,CAAC;IACf,yDAAyD;IACzD,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,gDAAgD,EAAE,UAAU,CAAC,CAAC;IAChF,oFAAoF;IACpF,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,gEAAgE,EAAE,QAAQ,CAAC,CAAC;IAC9F,+DAA+D;IAC/D,0EAA0E;IAC1E,+CAA+C;IAC/C,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,0FAA0F,EAAE,QAAQ,CAAC,CAAC;IACxH,0DAA0D;IAC1D,GAAG,GAAG,GAAG,CAAC,OAAO,CAAC,mEAAmE,EAAE,QAAQ,CAAC,CAAC;IACjG,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,WAAW,CAAC,IAAY;IACtC,MAAM,IAAI,GAAG,OAAO,EAAE,CAAC;IACvB,IAAI,GAAG,GAAG,gBAAgB,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IACvC,GAAG,GAAG,wBAAwB,CAAC,GAAG,CAAC,CAAC;IACpC,GAAG,GAAG,qBAAqB,CAAC,GAAG,CAAC,CAAC;IACjC,OAAO,GAAG,CAAC;AACb,CAAC;AAED,4DAA4D;AAC5D,MAAM,UAAU,SAAS,CAAC,IAAY;IACpC,IAAI,GAAG,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;IAC1B,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC;IACtB,GAAG,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC;IAC5B,GAAG,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC;IAC1B,GAAG,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC;IACrB,OAAO,GAAG,CAAC;AACb,CAAC;AAED,iFAAiF;AACjF,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,OAAO,mBAAmB,CAAC,IAAI,CAAC,CAAC;AACnC,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,UAAU,MAAM,CAAC,IAAY;IACjC,MAAM,IAAI,GAAG,OAAO,EAAE,CAAC;IACvB,IAAI,GAAG,GAAG,gBAAgB,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IACvC,GAAG,GAAG,wBAAwB,CAAC,GAAG,CAAC,CAAC;IACpC,GAAG,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC;IACrB,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC;IACtB,GAAG,GAAG,gBAAgB,CAAC,GAAG,CAAC,CAAC;IAC5B,GAAG,GAAG,mBAAmB,CAAC,GAAG,CAAC,CAAC;IAC/B,GAAG,GAAG,qBAAqB,CAAC,GAAG,CAAC,CAAC;IACjC,GAAG,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC;IAC1B,GAAG,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC;IACrB,OAAO,GAAG,CAAC;AACb,CAAC"}