feat: integrate learning-sdk into slackbot for persistent memory

[zzstoatzz]

Summary

• Add letta_api_key_secret_name setting to load Letta API key from Prefect secrets
• Wrap agent run with learning() context manager for persistent memory
• Add agentic-learning dependency to slackbot

Each Slack user gets their own memory agent (slackbot-{user_id}), enabling cross-session learning.

Test plan

• Create Prefect secret block letta-api-key with Letta API key
• Deploy slackbot and verify memory persists across conversations
• Test that different users have separate memory contexts

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR integrates the agentic-learning SDK into the slackbot to enable persistent memory across conversations. Each Slack user gets their own memory agent identified by slackbot-{user_id}, allowing the bot to remember context from previous interactions.

Key changes:

• Added optional Letta API key configuration via Prefect secrets
• Wrapped agent runs with learning() context manager for memory persistence
• Added agentic-learning as a dependency

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
examples/slackbot/src/slackbot/settings.py	Adds letta_api_key_secret_name setting and loads Letta API key from Prefect secrets in post-validation hook
examples/slackbot/src/slackbot/api.py	Imports learning from agentic_learning and wraps agent run with learning context manager using user-specific agent names
examples/slackbot/pyproject.toml	Adds agentic-learning to project dependencies

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

[nitpick] The broad exception handling with silent failure means that if the Letta API key secret fails to load for any reason (not just "doesn't exist"), the error will be suppressed. This is consistent with the turbopuffer pattern above (lines 95-100), but it makes debugging issues with secret configuration more difficult. Consider whether specific exception types should be caught (e.g., only catching the "secret not found" exception) to avoid masking other configuration errors.

Suggested change

	except Exception:
	pass # If secret doesn't exist, turbopuffer will handle the error
	if not os.getenv("LETTA_API_KEY"):
	try:
	api_key = Secret.load(self.letta_api_key_secret_name, _sync=True).get() # type: ignore
	os.environ["LETTA_API_KEY"] = api_key
	except Exception:
	except ValueError:
	pass # If secret doesn't exist, turbopuffer will handle the error
	if not os.getenv("LETTA_API_KEY"):
	try:
	api_key = Secret.load(self.letta_api_key_secret_name, _sync=True).get() # type: ignore
	os.environ["LETTA_API_KEY"] = api_key
	except ValueError: