Sync learning docs with current Learning Store APIs

[uzaxirr]

Summary

• update Learning Modes tool names and defaults to match current store behavior
• fix User Profile agentic tool name (update_profile)
• fix User Memory tool name, data model table, API example, and context tag
• clarify Decision Log default mode behavior in the store guide

[Copilot]

Pull request overview

Updates the Learning documentation to match current Learning Store API/tool naming and default-mode behavior, reducing drift between guides and runtime behavior.

Changes:

• Renames User Profile and User Memory agentic tool references to the current tool names (update_profile, update_user_memory).
• Updates User Memory data model + retrieval example and adjusts the injected context tag (<user_memory>).
• Clarifies Decision Log default-mode behavior in both the store guide and the Learning Modes overview.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

File	Description
learning/stores/user-profile.mdx	Updates agentic tool name and tightens User Profile vs User Memory guidance formatting.
learning/stores/user-memory.mdx	Updates tool name, data model table, retrieval snippet, and context-injection tag.
learning/stores/decision-log.mdx	Clarifies default mode behavior and annotates DecisionLogConfig() default.
learning/learning-modes.mdx	Updates tool table and default-mode guidance to match current store behavior.

Comments suppressed due to low confidence (1)

learning/stores/decision-log.mdx:33

• In this “Basic Usage” snippet, DecisionLogConfig() is now explicitly documented as defaulting to LearningMode.ALWAYS, but the agent instructions still tell the model to use the log_decision tool. Since this page later states tool-based logging is for Agentic mode, this example is internally inconsistent. Either switch the snippet to DecisionLogConfig(mode=LearningMode.AGENTIC) (and import LearningMode), or adjust the instructions/text to match ALWAYS mode behavior (automatic logging without relying on log_decision).

    learning=LearningMachine(
        decision_log=DecisionLogConfig(),  # defaults to LearningMode.ALWAYS
    ),
    instructions=[
        "When you make a significant choice, use log_decision to record it.",
        "Include your reasoning and alternatives you considered.",

---

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

[Copilot]

Only a single tool is listed here (update_user_memory), but the label says “Available tools” (plural). Consider changing to “Available tool” or briefly noting that update_user_memory covers add/update/delete/clear operations, to reduce ambiguity for readers.

Suggested change

	Available tools: `update_user_memory`
	Available tool: `update_user_memory` (supports add, update, delete, and clear operations)

[Copilot]

Docs now reference the update_profile tool, but there are still pages/examples in the repo that mention the old update_user_profile name (e.g. examples/learning/basics/b-user-profile-agentic.mdx and examples/learning/user-profile/agentic-mode.mdx). To avoid readers copying a stale tool name, please update the remaining references or add a brief note that update_user_profile was renamed/aliased to update_profile.

[Copilot]

The updated defaults table now says Entity Memory defaults to Always, but later in this same page the “Choosing a Mode” section still recommends Agentic for building entity knowledge graphs. If Entity Memory’s default/recommended behavior has changed, consider aligning the guidance below (and the earlier “Best for: … Entity Memory …” line) or clarifying that the default is Always but Agentic is recommended for certain workflows.

[shaloo]

@uzaxirr can you clarify the above as suggested?

[uzaxirr]

The Entity Memory default is Always — confirmed from the source code. The old docs had it as "Agentic" which was incorrect. I've pushed a fix that