v1.0.18 — list_memories, Session Briefing, tag_memory, inline related

What's new

list_memories tool

Browse all stored memories without needing a query. Filter by group, sort by date or name, paginate with cursors. Returns URI, name, group, timestamps, priority, confidence, and an excerpt per item. Fills the gap that search_memory left: now the AI can audit the archive or do a session warm-up without knowing what to search for.

Session Briefing resource

A virtual read-only resource pinned in every `resources/list` response at `pressocampus://[host]/briefing`. Reading it generates a fresh Markdown document with:

• Memory count and group breakdown
• Soul status and last-updated time
• Critical memories with excerpts
• Recent activity (last 7 days)
• Stale candidates (not updated in 6+ months, for review)

The soul is static identity; the briefing is dynamic session context. They serve different purposes.

tag_memory tool

Change an existing memory's group and/or priority without touching its content. Move orphaned memories into groups, promote key items to critical, merge scattered memories into a coherent structure. Protected resources (soul, index) cannot be retagged.

Inline related content in resources/read

When a memory has related URIs, the read response now includes annotations.related_content — an array of {uri, name, excerpt} for each linked memory. The AI no longer needs N extra round-trips to follow the relationship graph.