What does this PR do?

Removes the Honcho auto-migration path from AIAgent init so a blank memory.provider stays opt-in. This stops stale HONCHO_API_KEY / HONCHO_BASE_URL fallback state from rewriting memory.provider: honcho back into config at startup after the user has removed it.

Related Issue

N/A

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✅ Tests (adding or improving test coverage)

Changes Made

• Removed the Honcho auto-migration block from run_agent.py so memory provider activation only follows explicit config.
• Added a regression test in tests/run_agent/test_memory_provider_init.py that verifies blank memory.provider does not auto-enable Honcho or write config even when Honcho fallback would appear enabled.

How to Test

1. Set memory.provider to an empty string in ~/.hermes/config.yaml.
2. Leave stale HONCHO_API_KEY or HONCHO_BASE_URL present in env / ~/.hermes/.env.
3. Start Hermes and confirm it does not rewrite memory.provider: honcho back into config.

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: Ubuntu 24.04 / WSL2

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

Screenshots / Logs

Focused tests:

• scripts/run_tests.sh tests/run_agent/test_memory_provider_init.py tests/agent/test_memory_provider.py -q
• Result: 61 passed in 4.92s

Full suite on this checkout:

• scripts/run_tests.sh -q
• Result: 18 failed, 12947 passed, 36 skipped, 187 warnings in 299.58s (0:04:59)
• The failures are in unrelated existing areas on this checkout (gateway approvals, DingTalk, Discord bot filter, API server defaults, web_server schema, concurrent interrupt tests, backup profile wrappers, streaming tool call accumulation, browser camofox config version, tirith marker behavior).