| Age | Commit message (Expand) | Author |
|---|---|---|
| 2026-08-12 | feat: revise sandbox usability around semantic policy and LPAC process trees | tanaka takashi |
| 2026-08-12 | fix: implement native Ollama thinking semantics | tanaka takashi |
| 2026-08-12 | feat: harden Python workflow and structured planning | tanaka takashi |
| 2026-08-11 | Constrain structured planning responses | tanaka takashi |
| 2026-08-11 | Require task-specific tool evidence | tanaka takashi |
| 2026-08-11 | Verify task-specific external effect evidence | tanaka takashi |
| 2026-08-11 | Preserve planned tool parameters and bound failures | tanaka takashi |
| 2026-08-11 | Complete runtime diagnostics and verification coverage | tanaka takashi |
| 2026-08-11 | Audit and align v1 runtime design | tanaka takashi |
| 2026-08-11 | Fix: AgentRuntime tool failure handling and PlannerService fallback graph per... | tanaka takashi |
| 2026-08-11 | Fix PlanAsync fallback and BeginTask TaskGraph conflict | tanaka takashi |
| 2026-08-11 | Implement P2/P3: Plan-Run flow separation, PlannerService, and Windows test f... | tanaka takashi |
 |
index : winagent-sandbox-mcp.git | |
| Windows-native MCP agent sandbox for Ollama with LPAC, NTFS ACL, Job Objects, and WFP. |
| summaryrefslogtreecommitdiff |