@critiq/rules 0.1.0 → 0.3.0

package/rules/typescript/ts.security.no-dynamic-execution.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Eval or dynamic code execution
   summary: Eval-like helpers, `vm` execution APIs, and string-evaluated timers should not execute dynamic code.
   rationale: Dynamic execution turns data into code, widens the attack surface, and bypasses normal control flow.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-94
+      title: Improper Control of Generation of Code
+    - kind: owasp
+      title: Code Injection
+      url: https://owasp.org/www-community/attacks/Code_Injection
   tags:
     - security
     - execution
@@ -32,3 +41,4 @@ emit:
     summary: "`${captures.issue.text}` executes dynamic code and should be replaced with a safer alternative."
   remediation:
     summary: Replace dynamic execution with explicit parsing, fixed dispatch tables, or normal function callbacks.
+

package/rules/typescript/ts.security.no-fs-readfile-sync-in-handler.rule.yaml

@@ -0,0 +1,46 @@
+apiVersion: critiq.dev/v1alpha1
+kind: Rule
+metadata:
+  id: ts.security.no-fs-readfile-sync-in-handler
+  title: Avoid synchronous file reads in HTTP handlers
+  summary: Request handlers should not call `readFileSync` or equivalent blocking file APIs.
+  rationale: Blocking disk I/O inside request handlers stalls the Node.js event loop and hurts availability under load.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-16
+      title: Configuration
+    - kind: owasp
+      title: Secure Configuration Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Secure_Configuration_Cheat_Sheet.html
+  tags:
+    - security
+    - express
+    - performance
+    - rules-catalog
+  stability: stable
+  appliesTo: block
+scope:
+  languages:
+    - typescript
+    - javascript
+match:
+  fact:
+    kind: security.readfile-sync-in-request-handler
+    bind: issue
+emit:
+  finding:
+    category: security.availability
+    severity: medium
+    confidence: 0.86
+    tags:
+      - security
+      - express
+      - filesystem
+  message:
+    title: Replace blocking file read `${captures.issue.text}` in this handler
+    summary: "`${captures.issue.text}` performs synchronous disk I/O on the request path."
+  remediation:
+    summary: Use `fs.promises.readFile`, streams, or preload static assets outside the hot request path.
+

package/rules/typescript/ts.security.no-global-native-reassignment.rule.yaml

@@ -0,0 +1,44 @@
+apiVersion: critiq.dev/v1alpha1
+kind: Rule
+metadata:
+  id: ts.security.no-global-native-reassignment
+  title: Do not reassign global native bindings
+  summary: Do not assign to global native bindings such as `Object`, `Array`, or `undefined`.
+  rationale: Reassigning global natives breaks language invariants and can disable security checks that rely on them.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-94
+      title: Improper Control of Generation of Code
+    - kind: owasp
+      title: Code Injection
+      url: https://owasp.org/www-community/attacks/Code_Injection
+  tags:
+    - security
+    - language
+    - rules-catalog
+  stability: stable
+  appliesTo: block
+scope:
+  languages:
+    - typescript
+    - javascript
+match:
+  fact:
+    kind: security.global-native-reassignment
+    bind: issue
+emit:
+  finding:
+    category: security.language
+    severity: high
+    confidence: 0.97
+    tags:
+      - security
+      - language
+  message:
+    title: Do not reassign global native bindings
+    summary: "`${captures.issue.text}` reassigns a global native binding and can break runtime guarantees."
+  remediation:
+    summary: Use local variables with distinct names instead of overwriting global natives.
+

package/rules/typescript/ts.security.no-innerhtml-assignment.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid unsafe `innerHTML` assignment
   summary: "`innerHTML` assignments should only use fixed or explicitly sanitized HTML."
   rationale: Direct HTML injection can allow untrusted or weakly reviewed content to execute in the browser.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-79
+      title: Cross-site Scripting (XSS)
+    - kind: owasp
+      title: Cross Site Scripting Prevention Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Cross_Site_Scripting_Prevention_Cheat_Sheet.html
   tags:
     - security
     - output-encoding
@@ -34,3 +43,4 @@ emit:
     summary: "`${captures.issue.text}` inserts non-literal, non-sanitized HTML into `innerHTML`."
   remediation:
     summary: Prefer text-only rendering APIs or assign only fixed or explicitly sanitized HTML.
+

package/rules/typescript/ts.security.no-javascript-url.rule.yaml

@@ -0,0 +1,44 @@
+apiVersion: critiq.dev/v1alpha1
+kind: Rule
+metadata:
+  id: ts.security.no-javascript-url
+  title: Avoid `javascript:` URLs
+  summary: Do not use `javascript:` URLs in string literals, template literals, or JSX link attributes.
+  rationale: "`javascript:` URLs execute attacker-controlled code when used as navigation targets."
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-79
+      title: Cross-site Scripting (XSS)
+    - kind: owasp
+      title: Cross Site Scripting Prevention Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Cross_Site_Scripting_Prevention_Cheat_Sheet.html
+  tags:
+    - security
+    - xss
+    - rules-catalog
+  stability: stable
+  appliesTo: block
+scope:
+  languages:
+    - typescript
+    - javascript
+match:
+  fact:
+    kind: security.javascript-url
+    bind: issue
+emit:
+  finding:
+    category: security.output-encoding
+    severity: high
+    confidence: 0.94
+    tags:
+      - security
+      - xss
+  message:
+    title: "Avoid `javascript:` URLs"
+    summary: "`${captures.issue.text}` uses a `javascript:` URL that can execute arbitrary script."
+  remediation:
+    summary: "Use safe HTTPS links, in-app handlers, or explicit event callbacks instead of `javascript:` URLs."
+

package/rules/typescript/ts.security.no-native-prototype-extension.rule.yaml

@@ -0,0 +1,44 @@
+apiVersion: critiq.dev/v1alpha1
+kind: Rule
+metadata:
+  id: ts.security.no-native-prototype-extension
+  title: Do not extend native prototypes
+  summary: Do not assign properties on built-in prototype objects such as `Array.prototype`.
+  rationale: Mutating native prototypes affects every consumer of that type and can introduce subtle security bugs across the runtime.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-94
+      title: Improper Control of Generation of Code
+    - kind: owasp
+      title: Code Injection
+      url: https://owasp.org/www-community/attacks/Code_Injection
+  tags:
+    - security
+    - language
+    - rules-catalog
+  stability: stable
+  appliesTo: block
+scope:
+  languages:
+    - typescript
+    - javascript
+match:
+  fact:
+    kind: security.native-prototype-extension
+    bind: issue
+emit:
+  finding:
+    category: security.language
+    severity: high
+    confidence: 0.96
+    tags:
+      - security
+      - language
+  message:
+    title: Do not extend native prototypes
+    summary: "`${captures.issue.text}` mutates a built-in prototype and can change behavior globally."
+  remediation:
+    summary: Use utility functions or wrapper types instead of modifying native prototypes.
+

package/rules/typescript/ts.security.no-sync-child-process-exec.rule.yaml

@@ -0,0 +1,45 @@
+apiVersion: critiq.dev/v1alpha1
+kind: Rule
+metadata:
+  id: ts.security.no-sync-child-process-exec
+  title: Avoid synchronous child process execution with dynamic commands
+  summary: execSync and spawnSync should not run commands built from variables or template strings.
+  rationale: Synchronous shell execution blocks the event loop and dynamic command strings are a common command-injection surface.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-78
+      title: OS Command Injection
+    - kind: owasp
+      title: OS Command Injection Defense Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/OS_Command_Injection_Defense_Cheat_Sheet.html
+  tags:
+    - security
+    - express
+    - child-process
+    - rules-catalog
+  stability: stable
+  appliesTo: block
+scope:
+  languages:
+    - typescript
+    - javascript
+match:
+  fact:
+    kind: security.sync-child-process-exec
+    bind: issue
+emit:
+  finding:
+    category: security.command-injection
+    severity: high
+    confidence: 0.88
+    tags:
+      - security
+      - child-process
+  message:
+    title: Avoid synchronous shell execution in `${captures.issue.text}`
+    summary: "`${captures.issue.text}` runs a child process synchronously with a non-literal command string."
+  remediation:
+    summary: Prefer async APIs with fixed command allowlists, or validate and normalize inputs before invoking shell commands.
+

package/rules/typescript/ts.security.no-throw-literal.rule.yaml

@@ -0,0 +1,44 @@
+apiVersion: critiq.dev/v1alpha1
+kind: Rule
+metadata:
+  id: ts.security.no-throw-literal
+  title: Throw `Error` objects instead of literals
+  summary: Only throw `Error` instances (or subclasses), not strings, numbers, or plain objects.
+  rationale: Throwing literals loses stack traces and makes error handling inconsistent across callers.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-16
+      title: Configuration
+    - kind: owasp
+      title: Secure Configuration Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Secure_Configuration_Cheat_Sheet.html
+  tags:
+    - security
+    - reliability
+    - rules-catalog
+  stability: stable
+  appliesTo: block
+scope:
+  languages:
+    - typescript
+    - javascript
+match:
+  fact:
+    kind: security.throw-literal
+    bind: issue
+emit:
+  finding:
+    category: security.reliability
+    severity: medium
+    confidence: 0.9
+    tags:
+      - security
+      - reliability
+  message:
+    title: Throw an `Error` instead of a literal
+    summary: "`${captures.issue.text}` throws a non-`Error` value that is harder to inspect and handle safely."
+  remediation:
+    summary: Throw `new Error(...)` or a typed error subclass with a clear message.
+

package/rules/typescript/ts.security.no-with-statement.rule.yaml

@@ -0,0 +1,44 @@
+apiVersion: critiq.dev/v1alpha1
+kind: Rule
+metadata:
+  id: ts.security.no-with-statement
+  title: Avoid `with` statements
+  summary: "`with` statements make binding resolution unpredictable and are disallowed in strict mode."
+  rationale: The `with` statement introduces dynamic scope and makes static analysis, optimization, and security review unreliable.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-94
+      title: Improper Control of Generation of Code
+    - kind: owasp
+      title: Code Injection
+      url: https://owasp.org/www-community/attacks/Code_Injection
+  tags:
+    - security
+    - language
+    - rules-catalog
+  stability: stable
+  appliesTo: block
+scope:
+  languages:
+    - typescript
+    - javascript
+match:
+  fact:
+    kind: security.with-statement
+    bind: issue
+emit:
+  finding:
+    category: security.language
+    severity: high
+    confidence: 0.98
+    tags:
+      - security
+      - language
+  message:
+    title: Avoid `with` statements
+    summary: "`${captures.issue.text}` uses a `with` statement that makes binding resolution unpredictable."
+  remediation:
+    summary: Replace the `with` block with explicit property access on a named object.
+

package/rules/typescript/ts.security.non-literal-fs-filename.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid attacker-controlled filesystem read paths
   summary: Direct filesystem read APIs should not consume request- or upload-controlled filenames.
   rationale: Dynamic read paths can expose unintended local files or bypass expected file-selection constraints.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-434
+      title: Unrestricted Upload of File with Dangerous Type
+    - kind: owasp
+      title: File Upload Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/File_Upload_Cheat_Sheet.html
   tags:
     - security
     - filesystem
@@ -34,3 +43,4 @@ emit:
     summary: "`${captures.issue.text}` reads from a filename derived from external input."
   remediation:
     summary: Resolve reads from a trusted allowlist or a validated server-controlled mapping instead of external filenames.
+

package/rules/typescript/ts.security.nuxt-public-runtime-secret.rule.yaml

@@ -7,6 +7,18 @@ metadata:
     Sensitive credentials must not be exposed through runtimeConfig.public, which is visible to client bundles.
   rationale: >-
     Nuxt exposes runtimeConfig.public to the browser; placing secret material there leaks API keys, database passwords, and signing material to every visitor.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-798
+      title: Use of Hard-coded Credentials
+    - kind: owasp
+      title: Secrets Management Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Secrets_Management_Cheat_Sheet.html
+    - kind: url
+      title: Node.js security best practices
+      url: https://nodejs.org/en/learn/getting-started/security-best-practices
   tags:
     - security
     - nuxt
@@ -36,3 +48,4 @@ emit:
   remediation:
     summary: >-
       Keep secrets in the private runtimeConfig tree (non-public) and expose only publishable identifiers to the client after reviewing Nuxt runtime config documentation.
+

package/rules/typescript/ts.security.observable-timing-discrepancy.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Use constant-time secret comparison
   summary: Secrets and tokens should not be compared with ordinary equality operators.
   rationale: Ordinary string comparison can leak timing differences that help attackers guess secret material.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-327
+      title: Use of a Broken or Risky Cryptographic Algorithm
+    - kind: owasp
+      title: Cryptographic Storage Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Cryptographic_Storage_Cheat_Sheet.html
   tags:
     - security
     - cryptography
@@ -35,3 +44,4 @@ emit:
   remediation:
     summary: Use a constant-time comparison helper such as `crypto.timingSafeEqual` for secrets, tokens, and password hashes.
 
+

package/rules/typescript/ts.security.open-redirect.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Open redirect via request-controlled target
   summary: Redirect and navigation sinks should not use request-controlled destinations without validation.
   rationale: Redirect targets that are derived from user input can send authenticated users to attacker-controlled destinations.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-601
+      title: URL Redirection to Untrusted Site
+    - kind: owasp
+      title: Unvalidated Redirects and Forwards Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Unvalidated_Redirects_and_Forwards_Cheat_Sheet.html
   tags:
     - security
     - input-validation
@@ -37,3 +46,4 @@ emit:
   remediation:
     summary: Normalize the target to an internal path or validate it against a trusted origin allowlist before redirecting.
 
+

package/rules/typescript/ts.security.permissive-allow-origin.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Do not allow every origin in CORS policy
   summary: CORS should not fall back to wildcard or implicit allow-all origin settings.
   rationale: Wildcard or implicit allow-all CORS policies expose authenticated browser responses across origins.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-942
+      title: Permissive Cross-domain Policy with Untrusted Domains
+    - kind: owasp
+      title: Cross-Origin Resource Sharing (CORS)
+      url: https://owasp.org/www-community/attacks/CORS_OriginHeaderScrutiny
   tags:
     - security
     - cors
@@ -34,3 +43,4 @@ emit:
     summary: "`${captures.issue.text}` allows any origin through wildcard or implicit CORS policy."
   remediation:
     summary: Configure an explicit allowlist or validated origin callback instead of allowing all origins.
+

package/rules/typescript/ts.security.permissive-file-permissions.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid permissive file modes
   summary: Files that can carry user or security data should not be created with world-accessible modes.
   rationale: Broad file permissions expose application data to local users or processes that should not read or modify it.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-732
+      title: Incorrect Permission Assignment for Critical Resource
+    - kind: owasp
+      title: File Permission
+      url: https://owasp.org/www-community/vulnerabilities/Improper_File_Permissions
   tags:
     - security
     - filesystem
@@ -35,3 +44,4 @@ emit:
   remediation:
     summary: Use the minimum required file mode and remove world-readable or world-writable bits.
 
+

package/rules/typescript/ts.security.postmessage-wildcard-origin.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid wildcard `postMessage` targets
   summary: "`postMessage` calls should not use `*` as the target origin when they carry application data."
   rationale: Wildcard targets allow the browser to deliver the message to any origin that receives the window reference.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-79
+      title: Cross-site Scripting (XSS)
+    - kind: owasp
+      title: Cross Site Scripting Prevention Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Cross_Site_Scripting_Prevention_Cheat_Sheet.html
   tags:
     - security
     - browser
@@ -34,3 +43,4 @@ emit:
     summary: "`${captures.issue.text}` uses `*` as the `postMessage` target origin."
   remediation:
     summary: Set the target origin to a strict expected origin instead of `*`.
+

package/rules/typescript/ts.security.predictable-token-generation.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid predictable token generation
   summary: Tokens, reset links, and session secrets should be generated from cryptographically strong randomness.
   rationale: Predictable token material makes it easier to guess reset links, invite codes, and session secrets.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-287
+      title: Improper Authentication
+    - kind: owasp
+      title: JSON Web Token Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/JSON_Web_Token_for_Java_Cheat_Sheet.html
   tags:
     - security
     - authentication
@@ -34,3 +43,4 @@ emit:
     summary: "`${captures.issue.text}` builds a token-like value from predictable inputs."
   remediation:
     summary: Generate the value with `crypto.randomBytes`, `crypto.randomUUID`, or an approved secure token source.
+

package/rules/typescript/ts.security.raw-html-using-user-input.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid raw HTML with request input
   summary: Request-derived values should not be interpolated into raw HTML strings.
   rationale: Raw HTML construction with request data is a common path to reflected and stored XSS.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-79
+      title: Cross-site Scripting (XSS)
+    - kind: owasp
+      title: Cross Site Scripting Prevention Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Cross_Site_Scripting_Prevention_Cheat_Sheet.html
   tags:
     - security
     - xss
@@ -35,3 +44,4 @@ emit:
   remediation:
     summary: Use framework escaping, a trusted sanitizer, or safe DOM APIs instead of raw HTML interpolation.
 
+

package/rules/typescript/ts.security.request-driven-array-index-access.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid request-driven array indexing without bounds checks
   summary: Arrays indexed with request-derived keys can read or write out-of-bounds entries.
   rationale: Attacker-controlled indexes bypass assumptions about array length and element initialization.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-20
+      title: Improper Input Validation
+    - kind: owasp
+      title: Input Validation Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Input_Validation_Cheat_Sheet.html
   tags:
     - security
     - correctness
@@ -31,3 +40,4 @@ emit:
     summary: "Array access uses a computed index derived from ${captures.issue.text}."
   remediation:
     summary: Parse and bound-check indexes, prefer maps keyed by stable identifiers, or avoid indexing arrays with request data.
+

package/rules/typescript/ts.security.sensitive-data-egress.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Sensitive data egress to third-party processors
   summary: Sensitive values should not be sent to external processors or outbound SDKs without minimization or redaction.
   rationale: Sending regulated or secret data to third-party services increases privacy exposure and creates downstream processor risk.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-200
+      title: Exposure of Sensitive Information to an Unauthorized Actor
+    - kind: owasp
+      title: Logging Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Logging_Cheat_Sheet.html
   tags:
     - security
     - privacy
@@ -35,3 +44,4 @@ emit:
     summary: "`${captures.issue.text}` sends normalized sensitive datatypes to an outbound processor or external service."
   remediation:
     summary: Minimize the payload, redact the sensitive fields, or route the data only to approved processors.
+

package/rules/typescript/ts.security.sensitive-data-in-exception.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid sensitive data in thrown errors
   summary: Exceptions and rejection payloads should not include raw secrets or personal data.
   rationale: Exception payloads often reach logs, APM tools, and client responses with less review than normal business data.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-200
+      title: Exposure of Sensitive Information to an Unauthorized Actor
+    - kind: owasp
+      title: Logging Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Logging_Cheat_Sheet.html
   tags:
     - security
     - privacy
@@ -35,3 +44,4 @@ emit:
   remediation:
     summary: Replace raw secrets and personal data with opaque identifiers or redacted summaries before throwing or rejecting.
 
+

package/rules/typescript/ts.security.sensitive-data-written-to-file.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Avoid writing sensitive data to files
   summary: Data exports and local file writes should not persist raw secrets or personal fields.
   rationale: Static files, exports, and backups are easy to copy or retain beyond their intended lifetime.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-200
+      title: Exposure of Sensitive Information to an Unauthorized Actor
+    - kind: owasp
+      title: Logging Cheat Sheet
+      url: https://cheatsheetseries.owasp.org/cheatsheets/Logging_Cheat_Sheet.html
   tags:
     - security
     - privacy
@@ -35,3 +44,4 @@ emit:
   remediation:
     summary: Redact, hash, or exclude the sensitive fields before writing the file.
 
+

package/rules/typescript/ts.security.ssrf.rule.yaml

@@ -5,6 +5,15 @@ metadata:
   title: Server-side request forgery
   summary: Outbound requests should not use attacker-controlled targets or private hosts.
   rationale: Request-controlled targets can force the server to call internal services, metadata endpoints, or attacker-controlled infrastructure.
+  detection:
+    kind: pattern
+  references:
+    - kind: cwe
+      id: CWE-918
+      title: Server-Side Request Forgery (SSRF)
+    - kind: owasp
+      title: Server-Side Request Forgery
+      url: https://owasp.org/www-community/attacks/Server_Side_Request_Forgery
   tags:
     - security
     - network
@@ -33,3 +42,4 @@ emit:
     summary: "`${captures.ssrfCall.text}` sends an outbound request to a request-controlled or private target."
   remediation:
     summary: Resolve URLs against a trusted allowlist, reject private hosts, and proxy outbound requests through a vetted server-side helper.
+