Human in the loop - WIP

[onmete]

Description

Human-in-the-Loop (HITL) Feature

What

Adds human-in-the-loop approval workflow for AI tool execution. Users can now approve, deny, or modify tool calls before they execute.

Why

Provides safety controls for agentic AI operations. Critical for production environments where AI-initiated actions (like applying YAML or modifying resources) should require human oversight.

How It Works

1. Streaming query with tool call: When the LLM requests a tool execution, OLS checks if approval is required
2. Approval request: If required, OLS streams an approval_required event containing tool name, arguments, and an approval ID
3. User decision: User submits approval via POST /v1/hitl/approve with decision: approve, deny, or modify
4. Resume: OLS resumes the paused stream and either executes the tool or skips it based on the decision

API

• POST /v1/hitl/approve - Submit approval decision
• GET /v1/hitl/pending/{conversation_id} - List pending approvals

Configuration

hitl_config:
  enabled: true
  approval_timeout: 300
  auto_approve_read_only: true

Type of change

• Refactor
• New feature
• Bug fix
• CVE fix
• Optimization
• Documentation Update
• Configuration Update
• Bump-up dependent library
• Bump-up library or tool used for development (does not change the final image)
• CI configuration change
• Konflux configuration change

Related Tickets & Documents

• Related Issue #
• Closes #

Checklist before requesting a review

• I have performed a self-review of my code.
• PR has passed all pre-merge test jobs.
• If it is a core feature, I have added thorough tests.

Testing

• Please provide detailed steps to perform tests related to this code change.
• How were the fix/results from this change verified? Please provide relevant screenshots or results.

[openshift-ci]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign xrajesh for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[blublinsky]

I think the cofiguration does not belong here. It is configuration for the specific MCP headers and should be added there

[blublinsky]

This is cool, but... How does the client knows to invoke this? We have a request/reply paradigm in both OLS and LCOre. So the client makes a request and while waiting for a reply should invoke this? Does not seem right

[blublinsky]

Overall Human in the loop does not seem quite appropriate for request-reply. It is typically used in fire and forget scenarios