Issue imported from tinyhumansai/openhuman#2136
Created at: unknown
Problem
When a runtime/profile hides tools or installs policy, operators need proof of what is visible, what is blocked, and whether any raw write-capable tool bypass remains. Today this requires reading logs and reconstructing tool state manually.
Generic use case
Users and maintainers can inspect a profile/runtime and see the active tool inventory, policy mode, hidden raw tools, MCP allowlists, audit health, and recent denial reasons.
Managed-runtime use case
A compiled runtime contract can show conformance: source hash, policy hash, generated capability tools, hidden raw write tools, adapter map, and audit/approval health.
Proposed shape
- Add core diagnostics RPC that returns active policy/tool visibility state.
- Add conformance checks for raw write-capable tools, MCP allowlists, and audit availability.
- Add lightweight UI surface under settings/developer or diagnostics.
- Keep output redacted and support copyable support bundle.
Acceptance criteria
- RPC reports visible tools with permission/category metadata.
- RPC reports policy mode and recent policy denials.
- Tests cover conformance pass/fail for hidden vs exposed write tools.
- UI renders diagnostics without exposing secrets.
Alignment
This is generic observability for OpenHuman policy work. It helps contract-driven runtimes, Composio debugging, MCP debugging, and security reviews.
Problem
When a runtime/profile hides tools or installs policy, operators need proof of what is visible, what is blocked, and whether any raw write-capable tool bypass remains. Today this requires reading logs and reconstructing tool state manually.
Generic use case
Users and maintainers can inspect a profile/runtime and see the active tool inventory, policy mode, hidden raw tools, MCP allowlists, audit health, and recent denial reasons.
Managed-runtime use case
A compiled runtime contract can show conformance: source hash, policy hash, generated capability tools, hidden raw write tools, adapter map, and audit/approval health.
Proposed shape
Acceptance criteria
Alignment
This is generic observability for OpenHuman policy work. It helps contract-driven runtimes, Composio debugging, MCP debugging, and security reviews.