WebSocket/SSE origin verification by mabry1985 · Pull Request #3 · protoLabsAI/protoAgent

mabry1985 · 2026-04-19T06:05:44Z

Summary

Verify origins on websocket and SSE connections to the A2A streaming endpoint.

Problem

Anyone who can reach the A2A endpoint can open an SSE connection and drain another session's events if they know or guess the task ID. OpenClaw had the same class of bug — no origin verification. A2A spec leaves this to implementations.

Approach

Read A2A_ALLOWED_ORIGINS env (comma-separated) at startup; default empty (allow all, with WARNING log)
When set: reject SSE/WebSocket requests whose `Ori...

Recovered automatically by Automaker post-agent hook

Summary by CodeRabbit

New Features
- Added origin verification for A2A streaming endpoints to enhance security controls.
Documentation
- Added A2A security environment variables documentation for authentication token and origin allowlisting configuration.
Tests
- Added comprehensive test coverage for origin verification functionality.

mabry1985 · 2026-04-19T06:05:47Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:05:52Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:05:55Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

coderabbitai · 2026-04-19T06:05:58Z

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro

Run ID: 38955520-3a3f-4cea-b0a6-275f131a0fe8

📥 Commits

Reviewing files that changed from the base of the PR and between b89e30f and 832d551.

📒 Files selected for processing (4)

.automaker-lock
a2a_handler.py
docs/reference/environment-variables.md
tests/test_a2a_handler.py

Walkthrough

Added origin-header verification for A2A streaming endpoints. The implementation reads A2A_ALLOWED_ORIGINS at route registration time, validates Origin headers against the allowlist, and rejects disallowed origins with HTTP 403. Documentation and comprehensive test coverage were added.

Changes

Cohort / File(s)	Summary
Origin Verification Implementation `a2a_handler.py`	Added `_check_origin(request)` function with environment-driven allowlist logic. Integrated checks into JSON-RPC (`message/stream`, `tasks/resubscribe`) and REST (`/message:stream`, `GET /tasks/{task_id}:subscribe`) streaming endpoints; raises `HTTPException(403)` for disallowed origins while allowing requests missing `Origin` header.
Documentation `docs/reference/environment-variables.md`	Added "A2A security" subsection documenting `A2A_AUTH_TOKEN` and `A2A_ALLOWED_ORIGINS` environment variables, including wildcard support, case-insensitive comparison, and default warning behaviors.
Test Coverage `tests/test_a2a_handler.py`	Added comprehensive "Origin verification" test suite covering disallowed origins (HTTP 403), allowed requests without `Origin` header, empty-string rejection, wildcard permissiveness, and startup warning validation for unset environment variables.
Configuration `.automaker-lock`	Updated feature ID and timestamp fields.

Sequence Diagram

sequenceDiagram
    participant Client
    participant Handler as A2A Handler
    participant OriginCheck as Origin Checker
    participant Response as Response Handler

    Client->>Handler: POST /a2a (message/stream)<br/>with Origin header
    Handler->>OriginCheck: _check_origin(request)
    alt Origin in A2A_ALLOWED_ORIGINS
        OriginCheck->>Handler: Allow
        Handler->>Response: Create StreamingResponse
        Response->>Client: HTTP 200 + Stream
    else Origin not in allowlist
        OriginCheck->>Handler: HTTPException(403)
        Handler->>Response: Return JSONResponse(403)
        Response->>Client: HTTP 403 + Error detail
    else No Origin header
        OriginCheck->>Handler: Allow (skip check)
        Handler->>Response: Create StreamingResponse
        Response->>Client: HTTP 200 + Stream
    end

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

✨ Finishing Touches

📝 Generate docstrings

Create stacked PR
Commit on current branch

🧪 Generate unit tests (beta)

Create PR with unit tests
Commit unit tests in branch feature/websocket/sse-origin-verification

⚔️ Resolve merge conflicts

Resolve merge conflict in branch feature/websocket/sse-origin-verification

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

mabry1985 · 2026-04-19T06:05:58Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:01Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:03Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:06Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:09Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:12Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:14Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:17Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:20Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:23Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:25Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:28Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:31Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:34Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:36Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:39Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:42Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:45Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:47Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:50Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:53Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:56Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:06:59Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:07:01Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:01Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:03Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:06Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:09Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:12Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:15Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:17Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:20Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:22Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:25Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:28Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:31Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:34Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:36Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:39Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:42Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:45Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:48Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:51Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:53Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:56Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:31:59Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:32:02Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:32:04Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:32:07Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

mabry1985 · 2026-04-19T06:32:10Z

This PR has merge conflicts with the base branch and cannot be auto-merged.\n\nClosing and re-queuing the feature to backlog so it will be re-cut from the current base branch on the next auto-mode cycle.

Addresses bug #3 from v0.2.0 smoke test: MemoryMiddleware was silently skipped when knowledge: false (the default config), so session memory never worked out of the box. - graph/agent.py: drop `and knowledge_store` from activation guard — memory middleware now activates whenever memory: true, regardless of knowledge store - graph/middleware/memory.py: knowledge_store is now optional (default None); guard knowledge-extraction block when store is None; add standalone prior_sessions injection via before_model when running without KnowledgeMiddleware (no double-injection: only fires when self._store is None) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Addresses bug #3 from v0.2.0 smoke test: MemoryMiddleware was silently skipped when knowledge: false (the default config), so session memory never worked out of the box. - graph/agent.py: drop `and knowledge_store` from activation guard — memory middleware now activates whenever memory: true, regardless of knowledge store - graph/middleware/memory.py: knowledge_store is now optional (default None); guard knowledge-extraction block when store is None; add standalone prior_sessions injection via before_model when running without KnowledgeMiddleware (no double-injection: only fires when self._store is None) Co-authored-by: Ava <ava@protolabs.ai> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

refactor: WebSocket/SSE origin verification

832d551

mabry1985 closed this Apr 19, 2026

mabry1985 mentioned this pull request Apr 20, 2026

fix(memory): decouple MemoryMiddleware from knowledge_store #18

Merged

protoquinn Bot mentioned this pull request Apr 20, 2026

fix(goap): remove or re-register auto-triage-sweep — orphaned bug_triage routing causes 100% failure cascade #114

Closed

Conversation

mabry1985 commented Apr 19, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Problem

Approach

Summary by CodeRabbit

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

coderabbitai Bot commented Apr 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review failed

Walkthrough

Changes

Sequence Diagram

Estimated code review effort

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026

Uh oh!

mabry1985 commented Apr 19, 2026 •

edited by coderabbitai Bot

Loading

coderabbitai Bot commented Apr 19, 2026 •

edited

Loading