WOLFIE Platform · System Feature Map

Lightweight HTML prototype of the orchestration console. Demonstrates layer popovers, channel roster, and the LUPO-style workflow prompts that coordinate 80+ AI agents without overwhelming the human visitor.

Save any research session into WHO/WHAT/WHERE/WHEN/WHY/HOW slots. Collections map to LUPOPEDIA headers, giving every workflow a structured knowledge snapshot that travels with the team.

WOLFIE orchestrates core specialists (LUPO, LILITH, METIS, MAAT, CURSOR, ROSE, CLAUDE) using orchestration playbooks like Reverse Twenty Question Brainstorm and Incremental Plan & QA Loop.

99+ cultural contexts and personas, powering THALIA’s humor pipeline and all cross-cultural translations. ROSE keeps agents aligned to audience slang, metaphor, and tone.

Tiered deployment path: Tier 1 (MySQL shared hosting), Tier 2 (Postgres + feature flags), Tier 3 (pgvector + orchestration services). Everything is designed to migrate cleanly.

Reusable orchestration templates (Delegated Research Relay, Parallel Expert Panel, etc.) that spell out agent order, data capture checkpoints, and delivery expectations.

Header system keeps channel metadata, session IDs, and audit breadcrumbs synchronized across Sales Syntax, LUPOPEDIA collections, and the orchestration planner.

All documentation ships as `.txt` / `.md.txt` with metadata headers (WHO/WHAT/WHERE/WHEN/WHY/HOW/HELP). No .git dependencies, no cloud requirements—ready for 1999-era tooling.

Real-time overview of open channels, assigned agents, feature flags, and audit alerts. Mirrors the Sales Syntax console view so admins see exactly what the visitor sees.

Nightly orphan checks, foreign-key dry-runs, and tier-aware validation jobs. Reports go into `integrity_jobs` and `integrity_findings` tables for AGAPE/MAAT review.

One-click export of every table to CSV (matching the `/data/csv/` specs). Supports manual diffing, regression tests, and offline backups on the “cheap phone” workflow.

Manage AI roster entries (persona, skill tags, availability, vendor API keys). Tracks BIGINT IDs end-to-end so every orchestration decision links back to the source agent.

Feature flags for FK enforcement, HUMOR mode, brainstorming mode, and orchestration tier. Defaults to secure, offline-first settings.

Keeps WHO/WHAT/WHERE/WHEN/WHY/HOW/HELP metadata aligned across collections, chat transcripts, and orchestration plans. Ensures every export contains the full context block.

Tracks daily token budgets, $3/month hosting limits, and task queues. ERIS + METIS analyze conflicts and feed prevention protocols back into the orchestration layer.

Documents how to migrate from Tier 1 (phpMyAdmin) to Tier 2 (phpPgAdmin) to LUPO Admin. Includes manual test scripts, rollback instructions, and AGAPE compliance checks.

BIGINT-only IDs, orphan-friendly FK strategy, nightly validators, and CSV mirrors. Schema is tuned for Sales Syntax channels, collections, and orchestration audit trails.

Mockups define the channel list, agent badges, layer pop-ups, and workflow selectors. Serves as the blueprint for the future LUPO Admin surface.

All folders carry README indices, metadata headers, and offline-first `.md.txt` logs. The AGAPE/GENESIS/MD headers stay mandatory across every deliverable.

Tier 2 adds Postgres + pgRouting for orchestration state graphs. Tier 3 folds in pgvector, Temporal jobs, and LUPO Admin live dashboards—with clean migration scripts at each step.