fix(telegram-e2e): harden mtproto reply matching [AI-assisted]

[artemgetmann]

Summary

• Problem: the shared Telegram MTProto waiter could miss real tester-bot replies in DM/private-topic flows because it only looked at classic reply metadata and gave weak timeout diagnostics.
• Why it matters: shared Telegram E2E tooling can report false negatives even when the tester bot actually replied, which wastes time during runtime/browser verification.
• What changed: scripts/telegram-e2e/userbot_wait.py now understands direct_messages_topic.topic_id, emits ignored-message diagnostics on timeout, and reports sender mismatches clearly.
• What did NOT change (scope boundary): this PR does not reintroduce consumer-only smoke scripts on main; it only hardens the shared MTProto waiter.

Change Type (select all)

• Bug fix
• Feature
• Refactor
• Docs
• Security hardening
• Chore/infra

Scope (select all touched areas)

• Gateway / orchestration
• Skills / tool execution
• Auth / tokens
• Memory / storage
• Integrations
• API / contracts
• UI / DX
• CI/CD / infra

Linked Issue/PR

• Closes #
• Related fix(browser): port upstream existing-session readiness and profile targeting [AI-assisted] #104

User-visible / Behavior Changes

• Shared Telegram MTProto waiting logic is more reliable for DM/private-topic tester-bot replies.
• Timeout failures now show the ignored recent messages and why they were skipped.

Security Impact (required)

• New permissions/capabilities? (No)
• Secrets/tokens handling changed? (No)
• New/changed network calls? (No)
• Command/tool execution surface changed? (No)
• Data access scope changed? (No)
• If any Yes, explain risk + mitigation:

Repro + Verification

Environment

• OS: macOS
• Runtime/container: local checkout tooling
• Model/provider: N/A
• Integration/channel (if any): Telegram MTProto helper
• Relevant config (redacted): none

Steps

1. Run python3 -m py_compile scripts/telegram-e2e/userbot_wait.py.
2. Compare the waiter logic against the proven consumer fix.
3. Confirm the branch contains only the shared waiter hardening, not the consumer-only smoke wrapper.

Expected

• The waiter understands DM/private-topic metadata and surfaces actionable timeout diagnostics.

Actual

• The script compiles cleanly and the shared matching logic now mirrors the proven consumer waiter fix.

Evidence

Attach at least one:

• Failing test/log before + passing after
• Trace/log snippets
• Screenshot/recording
• Perf numbers (if relevant)

Human Verification (required)

What you personally verified (not just CI), and how:

• Verified scenarios:
  • python3 -m py_compile scripts/telegram-e2e/userbot_wait.py
  • inspected the main branch Telegram E2E layout and intentionally kept the consumer-only smoke wrapper out of this port
  • aligned the waiter behavior with the already proven consumer fix
• Edge cases checked:
  • direct_messages_topic.topic_id now resolves thread anchors
  • ignored-message diagnostics include sender/thread/text mismatch reasons
• What you did not verify:
  • a full live Telegram roundtrip from the main branch checkout
  • consumer-only smoke scripts, because they are not present on main

Review Conversations

• I replied to or resolved every bot review conversation I addressed in this PR.
• I left unresolved only the conversations that still need reviewer or maintainer judgment.

If a bot review conversation is addressed by this PR, resolve that conversation yourself. Do not leave bot review conversation cleanup for maintainers.

Compatibility / Migration

• Backward compatible? (Yes)
• Config/env changes? (No)
• Migration needed? (No)
• If yes, exact upgrade steps:

Failure Recovery (if this breaks)

• How to disable/revert this change quickly: revert commit 697e41ba2e
• Files/config to restore: scripts/telegram-e2e/userbot_wait.py
• Known bad symptoms reviewers should watch for: MTProto waiter timing out on real bot replies or reporting unhelpful timeout payloads

Risks and Mitigations

• Risk: Telegram may emit thread metadata differently across chat types.
  • Mitigation: the waiter now checks both classic reply metadata and direct_messages_topic.topic_id, and exposes ignored-message diagnostics when matching still fails.

AI Assistance

• AI-assisted
• Testing: lightly tested on main, fully proven on the equivalent consumer lane
• codex review was attempted, but local skill/MCP startup errors prevented a clean run in this environment