diff --git a/src/abi/calldata/decoder.ts b/src/abi/calldata/decoder.ts
index 6b7680b..6671833 100644
--- a/src/abi/calldata/decoder.ts
+++ b/src/abi/calldata/decoder.ts
@@ -2,19 +2,72 @@ import type {CalldataEncodable} from "../../types/calldata";
 import {CalldataAddress} from "../../types/calldata";
 import * as consts from "./consts";
 
+const fatalUtf8Decoder = new TextDecoder("utf-8", {fatal: true});
+
+function readByte(data: Uint8Array, index: {i: number}): number {
+  if (index.i >= data.length) {
+    throw new Error("unexpected end of calldata");
+  }
+  return data[index.i++];
+}
+
+function readBytes(
+  data: Uint8Array,
+  index: {i: number},
+  length: bigint | number,
+): Uint8Array {
+  const lengthBigInt = typeof length === "bigint" ? length : BigInt(length);
+  const remaining = BigInt(data.length - index.i);
+  if (lengthBigInt > remaining) {
+    throw new Error("unexpected end of calldata");
+  }
+
+  const lengthNumber = Number(lengthBigInt);
+  const ret = data.slice(index.i, index.i + lengthNumber);
+  index.i += lengthNumber;
+  return ret;
+}
+
 function readULeb128(data: Uint8Array, index: {i: number}): bigint {
-  let res: bigint = 0n;
+  let res = 0n;
   let accum = 0n;
-  let shouldContinue = true;
-  while (shouldContinue) {
-    const byte = data[index.i];
-    index.i++;
+
+  while (true) {
+    const byte = readByte(data, index);
     const rest = byte & 0x7f;
     res += BigInt(rest) * (1n << accum);
+
+    if ((byte & 0x80) === 0) {
+      if (byte === 0 && accum !== 0n) {
+        throw new Error("most significant ULEB128 octet cannot be zero");
+      }
+      return res;
+    }
+
     accum += 7n;
-    shouldContinue = byte >= 128;
   }
-  return res;
+}
+
+function compareCodePoints(left: string, right: string): number {
+  const leftPoints = Array.from(left, (char) => char.codePointAt(0)!);
+  const rightPoints = Array.from(right, (char) => char.codePointAt(0)!);
+
+  for (let i = 0; i < leftPoints.length && i < rightPoints.length; i++) {
+    const diff = leftPoints[i] - rightPoints[i];
+    if (diff !== 0) {
+      return diff;
+    }
+  }
+
+  return leftPoints.length - rightPoints.length;
+}
+
+function decodeUtf8(data: Uint8Array): string {
+  try {
+    return fatalUtf8Decoder.decode(data);
+  } catch {
+    throw new Error("invalid UTF-8 in calldata");
+  }
 }
 
 function decodeImpl(data: Uint8Array, index: {i: number}): CalldataEncodable {
@@ -27,28 +80,22 @@ function decodeImpl(data: Uint8Array, index: {i: number}): CalldataEncodable {
     case BigInt(consts.SPECIAL_FALSE):
       return false;
     case BigInt(consts.SPECIAL_ADDR): {
-      const res = data.slice(index.i, index.i + 20);
-      index.i += 20;
-      return new CalldataAddress(res);
+      return new CalldataAddress(readBytes(data, index, 20));
     }
   }
+
   const type = Number(cur & 0xffn) & ((1 << consts.BITS_IN_TYPE) - 1);
   const rest = cur >> BigInt(consts.BITS_IN_TYPE);
+
   switch (type) {
-    case consts.TYPE_BYTES: {
-      const ret = data.slice(index.i, index.i + Number(rest));
-      index.i += Number(rest);
-      return ret;
-    }
+    case consts.TYPE_BYTES:
+      return readBytes(data, index, rest);
     case consts.TYPE_PINT:
       return rest;
     case consts.TYPE_NINT:
       return -1n - rest;
-    case consts.TYPE_STR: {
-      const ret = data.slice(index.i, index.i + Number(rest));
-      index.i += Number(rest);
-      return new TextDecoder("utf-8").decode(ret);
-    }
+    case consts.TYPE_STR:
+      return decodeUtf8(readBytes(data, index, rest));
     case consts.TYPE_ARR: {
       const ret = [] as CalldataEncodable[];
       let elems = rest;
@@ -61,18 +108,31 @@ function decodeImpl(data: Uint8Array, index: {i: number}): CalldataEncodable {
     case consts.TYPE_MAP: {
       const ret = new Map<string, CalldataEncodable>();
       let elems = rest;
+      let previousKey: string | undefined;
+
       while (elems > 0) {
         elems--;
-        const strLen = Number(readULeb128(data, index));
-        const key = data.slice(index.i, index.i + strLen);
-        index.i += strLen;
-        const keyStr = new TextDecoder("utf-8").decode(key);
+        const strLen = readULeb128(data, index);
+        const keyStr = decodeUtf8(readBytes(data, index, strLen));
+
+        if (
+          previousKey !== undefined &&
+          compareCodePoints(previousKey, keyStr) >= 0
+        ) {
+          throw new Error(
+            `unordered calldata keys: '${previousKey}' >= '${keyStr}'`,
+          );
+        }
+
+        previousKey = keyStr;
         ret.set(keyStr, decodeImpl(data, index));
       }
       return ret;
     }
     default:
-      throw new Error(`can't decode type from ${type} rest is ${rest} at pos ${index.i}`);
+      throw new Error(
+        `can't decode type from ${type} rest is ${rest} at pos ${index.i}`,
+      );
   }
 }
 
diff --git a/tests/calldata.test.ts b/tests/calldata.test.ts
index 9b798bc..fb4c7e3 100644
--- a/tests/calldata.test.ts
+++ b/tests/calldata.test.ts
@@ -33,3 +33,80 @@ describe("calldata method-call encoding", () => {
     expect(decoded.has("kwargs")).toBe(true);
   });
 });
+
+describe("calldata decoder canonicality", () => {
+  it("rejects a truncated ULEB128 instead of reading past the input", () => {
+    expect(() => calldata.decode(new Uint8Array([0x80]))).toThrow(
+      "unexpected end of calldata",
+    );
+  });
+
+  it("rejects overlong ULEB128 encodings", () => {
+    expect(() => calldata.decode(new Uint8Array([0x80, 0x00]))).toThrow(
+      "most significant ULEB128 octet cannot be zero",
+    );
+  });
+
+  it.each([
+    ["address", new Uint8Array([0x18])],
+    ["bytes", new Uint8Array([0x0b])],
+    ["string", new Uint8Array([0x0c])],
+    ["array element", new Uint8Array([0x0d])],
+    ["map entry", new Uint8Array([0x0e])],
+  ])("rejects truncated %s payloads immediately", (_name, encoded) => {
+    expect(() => calldata.decode(encoded)).toThrow("unexpected end of calldata");
+  });
+
+  it("rejects invalid UTF-8 strings", () => {
+    expect(() => calldata.decode(new Uint8Array([0x0c, 0xff]))).toThrow(
+      "invalid UTF-8 in calldata",
+    );
+  });
+
+  it("rejects invalid UTF-8 map keys", () => {
+    expect(() =>
+      calldata.decode(new Uint8Array([0x0e, 0x01, 0xff, 0x00])),
+    ).toThrow("invalid UTF-8 in calldata");
+  });
+
+  it("rejects unordered map keys", () => {
+    const encoded = new Uint8Array([
+      0x16,
+      0x01,
+      0x62,
+      0x00,
+      0x01,
+      0x61,
+      0x00,
+    ]);
+
+    expect(() => calldata.decode(encoded)).toThrow(
+      "unordered calldata keys: 'b' >= 'a'",
+    );
+  });
+
+  it("rejects duplicate map keys", () => {
+    const encoded = new Uint8Array([
+      0x16,
+      0x01,
+      0x61,
+      0x00,
+      0x01,
+      0x61,
+      0x00,
+    ]);
+
+    expect(() => calldata.decode(encoded)).toThrow(
+      "unordered calldata keys: 'a' >= 'a'",
+    );
+  });
+
+  it("still round-trips canonical nested calldata", () => {
+    const input = new Map<string, CalldataEncodable>([
+      ["alpha", [1n, "hello", new Uint8Array([0x00, 0xff])]],
+      ["emoji", "😀"],
+    ]);
+
+    expect(calldata.decode(calldata.encode(input))).toEqual(input);
+  });
+});