feat(query): add overview() for project-aware session/memory discovery

Resolves current project from cwd, lists all known projects with session
  and memory counts, and returns the current project's recent sessions and
  memories in one call. Enables the agent to orient itself at the start of
  a retrieval without multiple exploratory queries.

references/retrieval-semantics.md

@@ -11,10 +11,16 @@ Classify the user's request before choosing tools.
 
 | User signal | Locator mode | Start with | Avoid first |
 |-------------|--------------|------------|-------------|
+| unclear project/session landscape | orientation | `overview()` | treating overview rows as evidence |
 | project name/path, session, cwd, file, time range | scope | `sessions()`, exact SQL on `project_path`, `sessionId`, `fileHistory()` | broad FTS |
 | workflow, subagent, tool call, summary, edit | artifact | `workflows()`, `subagents()`, `summaries()`, `tool_calls`, `tool_results` | all-session search |
 | concept, conclusion, design history, vague memory | semantic | `memories({ query })`, `search()`, summaries, bounded facet sweep | session dumps |
 
+`overview()` is a navigation map: current cwd/project if knowable, global
+project counts, and recent current-project session/memory entry points. Use it
+when scope is unclear, then query the memory or raw session layer for evidence.
+It does not guess the current session.
+
 One-shot retrieval is not all-shot retrieval. A query script may perform
 multiple steps, but the first locator should be the narrowest semantic fit. If a
 scope locator finds the relevant project/session/file, do not also run broad FTS
@@ -94,6 +100,15 @@ For semantic questions, build a task-local evidence view:
 Then synthesize the conclusion in the final answer. Do not pretend the raw
 evidence view is itself a stored Obelisk entity.
 
+After synthesis, check whether the conclusion should become a memory. Offer to
+write one when the result is durable, likely to help future sessions, and not
+already covered by `prior_memories`. Good candidates include design decisions,
+project conventions, abandoned alternatives, repeated failure causes, workflow
+patterns, and conclusions synthesized across multiple raw evidence points. Do
+not propose memory for one-off lookups, uncertain findings, or duplicate
+coverage. The offer is only a proposal: write the markdown file and run
+`--remember` only after user approval.
+
 ## Text Search Semantics
 
 `search(text)` passes text to SQLite FTS5 `MATCH`.