[MILESTONE] MemPalace × Evennia — Fleet Memory #1075
Closed
opened 2026-04-07 13:38:41 +00:00 by Timmy
·
6 comments
No Branch/Tag Specified
main
groq/issue-1126
groq/issue-1118
groq/issue-1119
claude/issue-1112
feat/mempalace-api-add-1775582323040
groq/issue-1047
groq/issue-915
claude/issue-1075
groq/issue-917
groq/issue-918
groq/issue-1103
groq/issue-1105
groq/issue-1106
groq/issue-1108
groq/issue-1092
groq/issue-1095
groq/issue-1098
groq/issue-913
timmy/issue-fix-896-897-898-910
claude/issue-823
claude/issue-879
claude/issue-880
claude/issue-827
claude/issue-882
claude/issue-826
claude/issue-836
claude/issue-832
claude/issue-833
timmy/issue-855
allegro/self-improvement-infra
ezra/deep-dive-architecture-20260405
claude/modularization-phase-1
gemini/issue-431
GoldenRockachopa
pre-agent-workers-v1
v0-golden
Labels
Clear labels
222-epic
3d-world
CI
QA
actionable
agent-presence
aistudio-ready
assigned-aistudio
assigned-claude
assigned-claw-code
assigned-gemini
assigned-groq
assigned-kimi
assigned-kimi
assigned-perplexity
assigned-sonnet
blocked
claude-ready
claw-code-done
claw-code-in-progress
deprioritized
duplicate
epic
gemini-api
gemini-review
google-ai-ultra
groq-ready
harness
identity
infrastructure
kimi-done
kimi-in-progress
kimi-ready
lazzyPit
media-gen
modularization
needs-design
nostr
p0-critical
p1-important
p2-backlog
performance
perplexity-ready
portal
research
security
sonnet-ready
sovereignty
velocity-engine
Continuous integration, runners, workflow issues
Quality assurance, testing, and production audit
Queued for Code Claw (qwen/openrouter)
Dispatched to Kimi via OpenClaw
Blocked by external dependency or merge conflict
Code Claw completed this task
Code Claw is actively working
Epic / umbrella issue
Gemini API integration
Google AI Ultra integration work
Timmy identity and branding
Kimi completed this task
Kimi is actively working on this
Lazarus Pit — automated agent resurrection and health recovery
AI media generation (image/video/audio)
Deep research and planning tasks
Security hardening, vulnerability fixes, access control
Auto-generated by velocity engine
No Label
Milestone
No items
No Milestone
MemPalace × Evennia — Fleet Memory
Projects
Clear projects
No project
Assignees
KimiClaw
Rockachopa
Timmy
allegro
antigravity
bezalel
claude
claw-code
codex-agent
ezra
gemini
google
grok
groq
hermes
kimi
manus
perplexity
sonnet
Clear assignees
No Assignees
claude
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: Timmy_Foundation/the-nexus#1075
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Meta tracking issue for the MemPalace × Evennia integration.
See spike: the-nexus #1073
Milestone: MemPalace × Evennia — Fleet Memory
Goal: Create a fleet-wide mind palace where every wizard persists knowledge in local vector rooms, and the Evennia world becomes the spatial, multiplayer interface to recall and act on that knowledge.
Sub-issues:
Velocity: 6 of 8 issues complete. 2 remaining (Phase 2 blocked on other wizards; Phase 3 security pending Alpha replication).
PR created: #1088
What was built
Full architectural scaffold for the MemPalace × Evennia fleet memory milestone.
nexus/mempalace/— Python APIconfig.py—MEMPALACE_PATH,FLEET_PALACE_PATH,FLEET_WING,CORE_ROOMS(env-var driven)searcher.py—search_memories(),search_fleet(),add_memory(),MemPalaceResult; ChromaDB lazy-imported withMemPalaceUnavailablefallbacknexus/evennia_mempalace/— Evennia pluginCommands (Phase 1 + Phase 4):
CmdRecall—recall <query>/recall <query> --fleet/recall <query> --room <room>CmdEnterRoom—enter room <topic>with keyword-aware room routingCmdRecord→hall_facts,CmdNote→hall_discoveries,CmdEvent→hall_eventsTypeclasses (Phase 1 + Phase 3):
MemPalaceRoom— description auto-refreshes from palace search on enterStewardNPC— interrogatable NPC that answers questions via palace searchdocs/mempalace/rooms.yaml— Taxonomy standard (#1082)5 required core rooms (
forge,hermes,nexus,issues,experiments), 3 write rooms, optional domain rooms, tunnel policy, privacy rules.Tests
40 new tests, all passing. ChromaDB fully mocked — tests run without a real palace.
PR updated: #1088
New deliverables added to claude/issue-1075:
#1082 — Taxonomy validation (new)
mempalace/validate_rooms.py— CI-ready script that checks a wizard palace against the fleet standard; exits non-zero on missing core rooms#1083 — Privacy boundary tools (new)
mempalace/export_closets.sh— closet-only export with hard abort on raw drawer or privatesource_fileviolationsmempalace/audit_privacy.py— weekly fleet palace scanner (raw drawers, full-text closets, private paths)Tests — 21 new pytest tests, all passing
Prior commit already covers:
nexus/evennia_mempalace/(full Evennia plugin with all Phase 1–4 commands + StewardNPC),docs/mempalace/rooms.yaml(comprehensive taxonomy with privacy policy), andnexus/mempalace/(Python API with ChromaDB searcher).Sub-issue status:
PR created: #1091
Deliverables in this round
#1079 (Agent NPCs): Added
CmdAsk— players can nowask <npc-name> about <topic>. The command searches for aStewardNPCin the current room (by name), falls back to location contents scan, and routes tonpc.respond_to_question(topic). Exported fromcommands/__init__.py. 3 new tests.#1080 (Live memory): Write commands (
CmdRecord/CmdNote/CmdEvent) now includeadded_by: evenniain stored metadata — satisfying the acceptance criterion. 1 new test.#1082 (Taxonomy CI): Added
docs/mempalace/bezalel_example.yaml(valid sample wizard config, all 5 core rooms) and avalidate-roomsCI step in.gitea/workflows/ci.ymlthat runsvalidate_rooms.pyagainst it. Validation script now passes in CI.Cleanup: Removed dead
mempalace/evennia_mempalace/duplicate (436 lines — superseded bynexus/evennia_mempalace/scaffold from prior merge).65 tests passing. Net: -347 lines.
Still blocked (infrastructure only):
PR created: #1109
Deliverables in this round
#1083 (Privacy boundary — weekly audit automation):
.gitea/workflows/weekly-audit.ymlwith aschedule: cron: "0 5 * * 1"trigger — the audit script now runs every Monday 05:00 UTC. Satisfies the "Audit script runs weekly and reports violations" acceptance criterion that was previously TBD.tests/fixtures/fleet_palace/with clean bezalel closet files (forge, hermes, issues rooms). The fixture passes the audit clean in CI, verifying the script and fixture together.#1078 (Shared fleet wing — Alpha-side access mechanism):
mempalace/fleet_api.py: lightweight JSON HTTP API (no extra deps) wrappingsearch_fleet(). ExposesGET /health,GET /search?q=…,GET /wings. HandlesMemPalaceUnavailablegracefully. Start withpython mempalace/fleet_api.pyon the Alpha host.12 new tests, all passing. 265 total.
Remaining (infrastructure-blocked):
Status update — #1075 milestone
All code-deliverable work is complete. PR #1109 is open with the final pieces.
What PR #1109 delivers
#1083 — Weekly audit automation (all acceptance criteria now met)
.gitea/workflows/weekly-audit.yml— runsaudit_privacy.pyon a Monday 05:00 UTC schedule against a CI fixturetests/fixtures/fleet_palace/— clean bezalel closet data (forge, hermes, issues rooms)#1078 — Alpha-side access mechanism (fleet HTTP API)
mempalace/fleet_api.py— stdlib JSON HTTP server:GET /health,GET /search?q=…,GET /wingsTest suite
test_evennia_ws_bridge.py)Remaining blocked on infrastructure only
These are operational blockers, not code blockers. The fleet API and audit cron are the last code artifacts.
PR created: #1110
Deliverables in this round
#1083 (Retention policy enforcement — final gap):
mempalace/retain_closets.py— 90-day closet aging enforcement. Removes*.closet.jsonfiles older than--days(default 90). Supports--dry-run. Added as a dry-run CI step inweekly-audit.ymlwith live cron guidance in comments.#1078 (Tunnel client — code side):
mempalace/tunnel_sync.py— stdlib-only HTTP client that pulls closets from a peerfleet_api.pyand writes them into the local fleet palace under the remote wing name. After sync,recall --fleetin Evennia returns results from the remote wizard wing. Supports--dry-run.fleet_api.py. The code side is complete.Tests: 29 new tests, all passing. Suite total: 294 passing.
Milestone status