Summary
Tool calls intermittently fail with [openclaw] missing tool result in session history; inserted synthetic error result for transcript repair. — the agent receives no output and the reply stalls until the user sends a follow-up message.

Steps to reproduce

1. Run OpenClaw with google-antigravity/claude-opus-4-5-thinking model via Telegram
2. Have an active conversation with multiple tool calls (e.g., exec, cron, gateway)
3. Observe that ~5-10% of tool calls return the synthetic error instead of actual results
4. Note: failures can cluster (2-3 consecutive failures observed)

Expected behavior
Tool calls should return their actual output, allowing the agent to continue the conversation.

Actual behavior
Tool calls sporadically return:

[openclaw] missing tool result in session history; inserted synthetic error result for transcript repair.The agent cannot proceed and goes silent. Retrying the same tool call (after user nudge) usually succeeds.

Environment
• OpenClaw version: 2026.2.2-3
• OS: Linux 6.14.0-37-generic (x64), Node v22.22.0
• Install method: npm (global)

Logs or screenshots
No errors appear in journalctl --user -u openclaw-gateway at the time of failure. The error message appears inline in the tool result payload. Frequency seems higher around gateway restarts but also occurs during normal operation.