Attacking Grpc Protobuf logo

Attacking Grpc Protobuf

Community
trilwu
attacking-grpc-protobuf

Test gRPC and Protocol Buffers services — recovering .proto definitions from server reflection or compiled descriptors, calling methods with grpcurl and grpcui, intercepting HTTP/2 and gRPC-Web traffic, and fuzzing unknown message schemas with protobuf-inspector. Use when a target speaks gRPC, HTTP/2 with application/grpc, or when a request body is opaque binary protobuf rather than JSON.

Overview

Publishertrilwu
Repositorysecskills
Skill nameattacking-grpc-protobuf
Stars
144
Forks
15
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by trilwu on GitHub. Read the source before you install it.

Installation

Install the Attacking Grpc Protobuf AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/trilwu/secskills.git /tmp/secskills
mkdir -p .claude/skills
cp -r /tmp/secskills/secskills-offense/skills/attacking-grpc-protobuf .claude/skills/attacking-grpc-protobuf
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Attacking Grpc Protobuf in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Attacking Grpc Protobuf on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Attacking Grpc Protobuf is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Attacking gRPC and Protobuf Services

gRPC breaks the tooling assumption that a request is readable text over HTTP/1.1. Burp shows a binary blob or nothing at all, and without the schema you cannot even name the methods. Recover the schema and it becomes an ordinary API test — with the twist that gRPC services are frequently internal services newly exposed, which means their authorization is often weaker than the REST API in front of them.

When to Use

  • Traffic uses HTTP/2 with content-type: application/grpc
  • A request or response body is opaque binary with no JSON structure
  • The app is a mobile or desktop client talking to a backend over protobuf
  • You find .proto files, *_pb2.py, *.pb.go, or grpc in a codebase
  • application/grpc-web or application/grpc-web+proto appears in a browser app

When NOT to Use

  • REST or GraphQL — use testing-apis
  • The proxy cannot see the traffic at all on mobile — use bypassing-mobile-pinning; gRPC clients often ignore system proxy settings
  • Source code is available — use auditing-code-for-vulnerabilities, and read the .proto files directly
  • Binary protocol that is not protobuf — use analyzing-binaries

Recover the Schema

Everything depends on this. Four routes, in order of cost.

1. Server reflection. Many services ship it enabled, often unintentionally.

bash
grpcurl -plaintext target:50051 list
grpcurl -plaintext target:50051 list package.ServiceName
grpcurl -plaintext target:50051 describe package.ServiceName.MethodName
grpcurl -plaintext target:50051 describe package.RequestMessage

# TLS
grpcurl target:443 list
grpcurl -insecure target:443 list          # self-signed / proxy in path

Reflection enabled on an internet-facing service is itself worth reporting — it is the gRPC equivalent of GraphQL introspection in production.

2. Descriptors compiled into a client. Protobuf embeds a serialized FileDescriptorProto in generated code, so the schema is recoverable from any client binary.

bash
# Mobile/desktop client: find and extract descriptor blobs
rg -a -o 'google/protobuf/descriptor.proto|\.proto' target_binary | head
# protobuf-inspector and protod can reconstruct .proto from embedded descriptors
protod target_binary -o ./protos

# JS/web clients: the descriptor is usually in the bundle as base64 or an array
rg -n 'grpc-web|serializeBinary|deserializeBinary' bundle.js | head

3. From the app's source or artifacts. .proto files in a repo, a Swagger gateway config, or generated stubs in a package.

4. Field-by-field inference. When you have neither, decode the wire format directly. Protobuf is self-describing enough to recover structure without the schema:

bash
protoc --decode_raw < message.bin
protobuf-inspector < message.bin
# Output: field numbers, wire types, and values — enough to fuzz and to
# recognize strings, nested messages, and integers

You lose field names but keep field numbers, which is all the wire format needs. That is sufficient to modify values and to add fields the client never sends.

Calling Methods

bash
# With reflection
grpcurl -plaintext -d '{"user_id": 1}' target:50051 package.Service/GetUser

# With a local .proto
grpcurl -import-path ./protos -proto api.proto \
        -d '{"user_id": 1}' target:50051 package.Service/GetUser

# Interactive browser UI, good for exploring
grpcui -plaintext target:50051

# Headers, including auth
grpcurl -H 'authorization: Bearer eyJ...' -plaintext target:50051 package.Service/List

For streaming methods, grpcurl accepts newline-delimited JSON on stdin for client streaming, and prints each message for server streaming. Streaming endpoints are frequently less-tested than unary ones and worth specific attention.

Intercepting Traffic

bash
# mitmproxy speaks HTTP/2 and can decode gRPC with a schema
mitmproxy --mode regular --set http2=true
#   the gRPC content-view renders protobuf; supply .proto for field names

# Burp: enable HTTP/2, and use a protobuf decoder extension
#   without one, you see length-prefixed binary frames

# gRPC-Web is easier — it rides HTTP/1.1 with base64 or binary framing,
# so a normal proxy sees the requests

gRPC framing: each message is a 1-byte compression flag, a 4-byte big-endian length, then the protobuf bytes. When a decoder shows nothing, strip those five bytes before feeding the payload to protoc --decode_raw.

bash
python3 -c "
import sys
d = sys.stdin.buffer.read()
sys.stdout.buffer.write(d[5:])" < frame.bin | protoc --decode_raw

Clients that ignore the system proxy need a network-layer redirect; see bypassing-mobile-pinning.

What to Test

The bug classes are the same as any API, but gRPC changes where they hide.

Authorization per method. gRPC has no path-based access control, so a reverse proxy or WAF that filters /admin/* does nothing. Each method must check authorization itself — enumerate every method from reflection and call each one with a low-privilege token.

bash
for m in $(grpcurl -plaintext target:50051 list package.Service | tail -n +2); do
  echo "== $m"; grpcurl -H "authorization: Bearer $LOW_PRIV" -plaintext -d '{}' target:50051 "$m" 2>&1 | head -3
done

Internal services exposed. gRPC is a service-mesh protocol, so many services were written assuming only other services would call them. If you can reach one directly, expect no authentication at all — and expect it to trust identity claims passed as ordinary request fields or metadata.

Metadata trust. Look for headers the service reads as identity: x-user-id, x-tenant-id, x-forwarded-user. If a gateway sets them and the service trusts them, sending them yourself is a complete authentication bypass.

bash
grpcurl -H 'x-user-id: 1' -H 'x-role: admin' -plaintext -d '{}' target:50051 package.Service/GetProfile

Unknown-field injection. Protobuf ignores fields it does not recognize, but intermediate services may forward them. More usefully: add fields the client never sends but the server schema defines — is_admin, internal_notes, tenant_id — the mass-assignment equivalent.

Type confusion and resource exhaustion. Wire types are loosely enforced; send a bytes where a string is expected, deeply nested messages to blow the recursion limit, or a huge repeated field. Check for a configured message size limit.

TLS and mTLS posture. Many gRPC deployments use insecure channels internally. Check whether the service accepts plaintext, and whether mTLS is required or merely optional.

Rationalizations to Reject

  • "Burp shows nothing, the app isn't making requests." Burp needs HTTP/2 enabled, and gRPC clients often bypass the system proxy entirely.
  • "I don't have the .proto, so I can't test it." protoc --decode_raw recovers structure from any message. Field numbers are all you need.
  • "Reflection is disabled, so the methods are hidden." The client knows the schema. Extract the descriptors from the client binary.
  • "The gateway enforces authorization." The gateway sees method names, not intent, and anything that reaches the service directly bypasses it entirely.
  • "It's internal-only." Confirm that. Service meshes leak, and "internal" usually means "no authentication".
  • "The client never sends that field." You are not the client.

ATT&CK Coverage

Generated from secskills-core/ttp-index.json — edit that file, then run python3 scripts/sync_attack.py --write. Re-verify IDs against the current ATT&CK release before citing them in a report.

Initial Access (TA0001)

  • T1190 Exploit Public-Facing Application — see also testing-web-applications, testing-apis, enumerating-network-services, attacking-graphql, exploiting-deserialization, exploiting-ssrf, exploiting-xxe

Detection content for any of these: engineering-detections. Proactive search: hunting-threats. Post-compromise: responding-to-incidents.

References

  • testing-apis — the general API methodology this specializes
  • auditing-code-for-vulnerabilities — reading .proto and handlers in source
  • bypassing-mobile-pinning — when a mobile gRPC client ignores your proxy
  • exploiting-cloud-platforms — service mesh and internal exposure context
  • grpcurl, grpcui, protoc, protobuf-inspector, protod, mitmproxy

Frequently asked questions

What does the Attacking Grpc Protobuf AI skill do?

Test gRPC and Protocol Buffers services — recovering .proto definitions from server reflection or compiled descriptors, calling methods with grpcurl and grpcui, intercepting HTTP/2 and gRPC-Web traffic, and fuzzing unknown message schemas with protobuf-inspector. Use when a target speaks gRPC, HTTP/2 with application/grpc, or when a request body is opaque binary protobuf rather than JSON.

Why use Attacking Grpc Protobuf on TypingMind?

Because you install it once and use it with any model. Attacking Grpc Protobuf is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Attacking Grpc Protobuf in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/trilwu/secskills/tree/main/secskills-offense/skills/attacking-grpc-protobuf. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Attacking Grpc Protobuf?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Attacking Grpc Protobuf?

As many as you like. As long as a model supports skills, you can use Attacking Grpc Protobuf with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Attacking Grpc Protobuf AI skill free?

Yes. It is published on GitHub by trilwu under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇