fix: prevent tools invocation without valid session initialization #607
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Modified InsecureStatefulSessionIdManager to track active sessions using sync.Map - Added session existence validation in addition to format validation - Implemented proper session termination tracking - Added comprehensive regression tests for session validation scenarios - Updated sampling tests to use stateless mode for compatibility This fixes a security issue where tools could be invoked with any well-formatted session ID without going through proper initialization, allowing unauthorized access.
WalkthroughImplements stateful session tracking in the HTTP server by adding internal maps for active and terminated sessions to InsecureStatefulSessionIdManager, updating Generate/Validate/Terminate logic. Updates server construction to accept a WithStateLess option. Expands tests to cover session lifecycle validation, termination behavior, and stateless construction in sampling tests. Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Possibly related PRs
Suggested reviewers
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Merging this branch will increase overall coverage
Coverage by fileChanged files (no unit tests)
Please note that the "Total", "Covered", and "Missed" counts above refer to code statements instead of lines of code. The value in brackets refers to the test coverage of that file in the old version of the code. Changed unit test files
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
server/streamable_http.go(1 hunks)server/streamable_http_sampling_test.go(3 hunks)server/streamable_http_test.go(1 hunks)
🧰 Additional context used
🧬 Code graph analysis (2)
server/streamable_http_test.go (4)
mcp/tools.go (3)
NewTool(679-701)CallToolRequest(54-58)CallToolResult(40-51)mcp/utils.go (1)
NewToolResultText(271-280)server/streamable_http.go (2)
NewTestStreamableHTTPServer(1060-1064)InsecureStatefulSessionIdManager(1021-1024)server/constants.go (1)
HeaderKeySessionID(5-5)
server/streamable_http_sampling_test.go (1)
server/streamable_http.go (2)
NewStreamableHTTPServer(152-167)WithStateLess(45-51)
Comment on lines
1050
to
1056
| func (s *InsecureStatefulSessionIdManager) Terminate(sessionID string) (isNotAllowed bool, err error) { | ||
| if _, exists := s.sessions.Load(sessionID); !exists { | ||
| return false, fmt.Errorf("session not found: %s", sessionID) | ||
| } | ||
| s.terminated.Store(sessionID, true) | ||
| s.sessions.Delete(sessionID) | ||
| return false, nil |
There was a problem hiding this comment.
Make session termination idempotent
Terminate now returns an error once the session has been deleted from sessions, so a client that retries DELETE /... after succeeding immediately hits a 500 (handleDelete turns this error into Session termination failed). Termination endpoints must be idempotent—network retries shouldn’t look like server failures. Please allow already-terminated IDs to short‑circuit as success (or map them to a 404) before returning an error.
func (s *InsecureStatefulSessionIdManager) Terminate(sessionID string) (isNotAllowed bool, err error) {
- if _, exists := s.sessions.Load(sessionID); !exists {
- return false, fmt.Errorf("session not found: %s", sessionID)
- }
- s.terminated.Store(sessionID, true)
- s.sessions.Delete(sessionID)
+ if _, exists := s.sessions.Load(sessionID); !exists {
+ if _, terminated := s.terminated.Load(sessionID); terminated {
+ return false, nil
+ }
+ return false, fmt.Errorf("session not found: %s", sessionID)
+ }
+ s.terminated.Store(sessionID, true)
+ s.sessions.Delete(sessionID)
return false, nil
}🤖 Prompt for AI Agents
In server/streamable_http.go around lines 1050–1056, Terminate currently returns
an error when the session ID is not found in s.sessions, causing retries to
surface as server failures; change the logic to make termination idempotent by
first checking whether the session is already marked terminated (or if not
found, consult s.terminated) and if so return (false, nil) to indicate success,
otherwise mark terminated and delete the session and return (false, nil); ensure
you do not treat a missing sessions entry as an internal error when the session
is already terminated.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
This PR fixes a security vulnerability where tools could be invoked without proper session initialization. The
InsecureStatefulSessionIdManagerwas only validating session ID format but not checking if the session actually existed, allowing any well-formatted fake session ID to bypass authentication.Fixes #579
Type of Change
Changes Made
Core Implementation
InsecureStatefulSessionIdManagerto track active sessions usingsync.Mapsessionsmap to store generated session IDsterminatedmap to track terminated sessionsGenerate()to store session IDs when createdValidate()to check session existence in addition to format validationTerminate()to properly mark sessions as terminated and remove from active sessionsTesting
TestStreamableHTTP_SessionValidationwith comprehensive test cases:TestInsecureStatefulSessionIdManagerfor unit testing the session manager:TestStreamableHTTPServer_SamplingErrorHandlingto use stateless mode for compatibilityChecklist
Additional Information
Security Impact
Before this fix, an attacker could invoke tools using any well-formatted session ID like
mcp-session-ffffffff-ffff-ffff-ffff-ffffffffffffwithout going through the initialization flow. This fix ensures that only session IDs that were actually generated by the server are accepted.Verification
The original exploit from issue #579 now properly returns a 400 Bad Request with "Invalid session ID" error message instead of executing the tool.
All existing tests pass, demonstrating backward compatibility with legitimate use cases.
Summary by CodeRabbit