Merge pull request #128 from Kzoeps/97-add-integrity-check-for-evidence-content-string-being-a-uri

feat(sdk-core): add URI validation for attachment content strings

Author: aspiers

.changeset/add-content-uri-validation.md

@@ -0,0 +1,23 @@
+---
+"@hypercerts-org/sdk-core": minor
+---
+
+**BREAKING CHANGE (pre-1.0):** Add strict URI validation for attachment content strings
+
+This introduces a breaking behavioral change for 0.x consumers:
+
+- Add `isValidUri` utility to validate URI strings have a proper scheme (supports http, https, at://, ipfs://, and any
+  RFC 3986-compliant scheme)
+- **`addAttachment` now throws `ValidationError`** when content strings are not valid URIs (e.g., plain text like
+  `"not-a-uri"`)
+- Export `isValidUri` from the public API for consumer use
+
+**Migration Guide:** Existing code that passes plain text or non-URI strings to `addAttachment` will now fail with a
+`ValidationError`. To migrate:
+
+1. Ensure all content strings passed to `addAttachment` are valid URIs
+2. Use the new `isValidUri` utility to validate strings before passing them
+3. Convert plain text content to proper URI format (e.g., data URIs, IPFS URIs, or HTTP URLs)
+
+**Compatibility Note:** Consumers relying on the previous lenient behavior that accepted non-URI strings must update
+their code. The validation now strictly enforces that attachment content must be a valid URI with a recognized scheme.

packages/sdk-core/src/index.ts

@@ -250,6 +250,9 @@ export type {
   Permission,
 } from "./auth/permissions.js";
 
+// URL Utilities
+export { isValidUri } from "./lib/url-utils.js";
+
 // Rich Text Utilities
 export { createFacetsFromText, createFacetsFromTextSync, RichText } from "./lib/rich-text.js";
 export type { RichTextResult, AppBskyRichtextFacet } from "./lib/rich-text.js";

packages/sdk-core/src/lib/url-utils.ts

@@ -1,3 +1,46 @@
+/**
+ * Regular expression to match a valid URI with a scheme.
+ *
+ * Matches strings that start with a scheme (one or more alphanumeric characters,
+ * plus, period, or hyphen) followed by a colon. This covers schemes that the
+ * native URL constructor may not support (e.g., `at://`, `ipfs://`).
+ *
+ * @see https://www.rfc-editor.org/rfc/rfc3986#section-3.1
+ * @internal
+ */
+const URI_SCHEME_REGEX = /^[a-zA-Z][a-zA-Z0-9+\-.]*:/;
+
+/**
+ * Check if a string is a valid URI with a scheme.
+ *
+ * Validates that the string is a properly formatted URI. Uses the native
+ * `URL` constructor for standard schemes (http, https, ftp, etc.) and falls
+ * back to scheme detection for non-standard schemes like `at://` and `ipfs://`
+ * that the `URL` constructor does not support.
+ *
+ * @param uri - The string to validate
+ * @returns True if the string is a valid URI with a scheme, false otherwise
+ *
+ * @example
+ * ```typescript
+ * isValidUri("https://example.com/report.pdf"); // true
+ * isValidUri("ipfs://QmYwAPJzv5CZsnA625s3Xf2nemtYgPpHdWEz79ojWnPbdG"); // true
+ * isValidUri("at://did:plc:abc/org.col/rkey"); // true
+ * isValidUri("not-a-uri"); // false
+ * isValidUri(""); // false
+ * ```
+ */
+export function isValidUri(uri: string): boolean {
+  if (!uri) return false;
+
+  try {
+    new URL(uri);
+    return true;
+  } catch {
+    return URI_SCHEME_REGEX.test(uri);
+  }
+}
+
 /**
  * Type guard to check if a URL is a loopback address.
  *

packages/sdk-core/src/repository/HypercertOperationsImpl.ts

@@ -54,6 +54,7 @@ import type { CreateResult, ListParams, PaginatedList, ProgressStep, UpdateResul
 import { uploadResultToBlobRef } from "./types.js";
 import { $Typed } from "@atproto/api";
 import { sha256Hash } from "../lib/crypto.js";
+import { isValidUri } from "../lib/url-utils.js";
 
 /**
  * Implementation of high-level hypercert operations.
@@ -1109,12 +1110,22 @@ export class HypercertOperationsImpl extends EventEmitter<HypercertEvents> imple
    *
    * @param contentInput - Single content item or array (URI strings or Blobs)
    * @returns Promise resolving to array of URI refs or Blob refs
+   * @throws {@link ValidationError} if a string content item is not a valid URI
    * @throws {@link NetworkError} if blob upload fails
    * @internal
    */
   private async resolveAttachmentContent(contentInput: string | Blob | Array<string | Blob>) {
     const contentArray = Array.isArray(contentInput) ? contentInput : [contentInput];
 
+    // Validate that all string content items are valid URIs before resolving
+    for (const item of contentArray) {
+      if (typeof item === "string" && !isValidUri(item)) {
+        throw new ValidationError(
+          `Invalid URI: "${item}". Content must be a valid URI with a scheme (e.g., https://example.com)`,
+        );
+      }
+    }
+
     return await Promise.all(contentArray.map((item) => this.resolveUriOrBlob(item, "application/octet-stream")));
   }
 

packages/sdk-core/tests/lib/url-utils.test.ts

@@ -0,0 +1,69 @@
+import { describe, it, expect } from "vitest";
+import { isValidUri } from "../../src/lib/url-utils.js";
+
+describe("isValidUri", () => {
+  describe("valid URIs", () => {
+    it("should accept https URLs", () => {
+      expect(isValidUri("https://example.com")).toBe(true);
+      expect(isValidUri("https://example.com/path/to/resource")).toBe(true);
+      expect(isValidUri("https://example.com/report.pdf")).toBe(true);
+      expect(isValidUri("https://example.com:8080/path?q=1&r=2#frag")).toBe(true);
+    });
+
+    it("should accept http URLs", () => {
+      expect(isValidUri("http://example.com")).toBe(true);
+      expect(isValidUri("http://localhost:3000")).toBe(true);
+    });
+
+    it("should accept AT Protocol URIs", () => {
+      expect(isValidUri("at://did:plc:abc123/org.hypercerts.claim.activity/rkey")).toBe(true);
+      expect(isValidUri("at://did:plc:test/org.hypercerts.claim.record/def456")).toBe(true);
+      expect(isValidUri("at://did:web:example.com/app.bsky.feed.post/3km2vj4kfqp2a")).toBe(true);
+    });
+
+    it("should accept ftp URIs", () => {
+      expect(isValidUri("ftp://files.example.com/doc.txt")).toBe(true);
+    });
+
+    it("should accept ipfs URIs", () => {
+      expect(isValidUri("ipfs://QmYwAPJzv5CZsnA625s3Xf2nemtYgPpHdWEz79ojWnPbdG")).toBe(true);
+      expect(isValidUri("ipfs://bafybeigdyrzt5sfp7udm7hu76uh7y26nf3efuylqabf3oclgtqy55fbzdi")).toBe(true);
+    });
+
+    it("should accept data URIs", () => {
+      expect(isValidUri("data:text/plain;base64,SGVsbG8=")).toBe(true);
+    });
+
+    it("should accept mailto URIs", () => {
+      expect(isValidUri("mailto:user@example.com")).toBe(true);
+    });
+  });
+
+  describe("invalid URIs", () => {
+    it("should reject plain text", () => {
+      expect(isValidUri("not-a-uri")).toBe(false);
+      expect(isValidUri("just some text")).toBe(false);
+      expect(isValidUri("hello world")).toBe(false);
+    });
+
+    it("should reject empty strings", () => {
+      expect(isValidUri("")).toBe(false);
+    });
+
+    it("should reject bare hostnames without scheme", () => {
+      expect(isValidUri("example.com")).toBe(false);
+      expect(isValidUri("www.example.com")).toBe(false);
+    });
+
+    it("should reject relative paths", () => {
+      expect(isValidUri("/relative/path")).toBe(false);
+      expect(isValidUri("./relative/path")).toBe(false);
+      expect(isValidUri("../parent/path")).toBe(false);
+    });
+
+    it("should reject strings with only a colon", () => {
+      expect(isValidUri(":not-valid")).toBe(false);
+      expect(isValidUri("://missing-scheme")).toBe(false);
+    });
+  });
+});

packages/sdk-core/tests/repository/HypercertOperationsImpl.test.ts

@@ -1464,6 +1464,53 @@ describe("HypercertOperationsImpl", () => {
         }),
       ).rejects.toThrow(NetworkError);
     });
+
+    it("should throw ValidationError when content URI is not a valid URI", async () => {
+      await expect(
+        hypercertOps.addAttachment({
+          subjects: "at://did:plc:test/org.hypercerts.claim.activity/abc",
+          title: "Attachment with invalid URI",
+          content: "not-a-valid-uri",
+        }),
+      ).rejects.toThrow(ValidationError);
+    });
+
+    it("should throw ValidationError when content URI is plain text", async () => {
+      await expect(
+        hypercertOps.addAttachment({
+          subjects: "at://did:plc:test/org.hypercerts.claim.activity/abc",
+          title: "Attachment with plain text",
+          content: "just some random text",
+        }),
+      ).rejects.toThrow(ValidationError);
+    });
+
+    it("should throw ValidationError when any content URI in array is invalid", async () => {
+      await expect(
+        hypercertOps.addAttachment({
+          subjects: "at://did:plc:test/org.hypercerts.claim.activity/abc",
+          title: "Attachment with mixed content",
+          content: ["https://example.com/valid.pdf", "not-a-valid-uri"],
+        }),
+      ).rejects.toThrow(ValidationError);
+    });
+
+    it("should accept valid non-http URI schemes in content", async () => {
+      const result = await hypercertOps.addAttachment({
+        subjects: "at://did:plc:test/org.hypercerts.claim.activity/abc",
+        title: "IPFS Attachment",
+        content: "ipfs://QmYwAPJzv5CZsnA625s3Xf2nemtYgPpHdWEz79ojWnPbdG",
+      });
+
+      expect(result.uri).toContain("attachment");
+      const call = mockAgent.com.atproto.repo.createRecord.mock.calls[0][0];
+      expect(call.record.content).toEqual([
+        {
+          $type: "org.hypercerts.defs#uri",
+          uri: "ipfs://QmYwAPJzv5CZsnA625s3Xf2nemtYgPpHdWEz79ojWnPbdG",
+        },
+      ]);
+    });
   });
 
   describe("addMeasurement", () => {