← Back to the capability index
AI · ai.agent-operationsAgent Operations (Validation Layer)
Partial coverage Pro
Evidence by type
| Evidence type | Detail | Result |
|---|---|---|
| No distinct route | cross-cutting-gate | Validation-layer gate applied across several existing MCP/spec routes rather than a distinct route of its own; see ai.spec-apply, ai.grounding, ai.workflow-generation for the gated route surfaces. |
SDK coverage
| SDK | Status | Entry points | Note |
|---|---|---|---|
| honua-sdk-js | partial | dryRunAgentPlan, executeAgentPlanStep, verifyAgentStepAuthorization (src/agent-safety) | Agent dry-run, execution, and approval-consumption authorization are implemented in src/agent-safety and proven by dedicated tests. The MCP tool-discovery/execution facade (agent-tools-facade in config/support-manifest.v1.json) additionally requires a Honua Server for tool execution against live data, so this key stops short of a fully standalone claim. |
| honua-sdk-dotnet | not-covered | — | — |
| honua-sdk-python | not-covered | — | — |
Executable samples
No executable sample published for this capability yet.
Known gaps
No open honua-server issues tagged for this capability’s category.
Raw receipts
- capability-keys.v1.json — canonical key definition for
ai.agent-operations. - capability-matrix.v1.json — honua-server's Phase-A evidence snapshot (entry/test counts, CITE, parity, interop, geobench).
- sdk-js coverage · sdk-dotnet coverage · sdk-python coverage
- honua-samples run-samples CI — executable sample run history.
- capability-matrix.v1.json — the full enriched aggregate backing this page.