Testing environment
Launch the political-strategy Interface Lab
Use the live testing environment to pressure-test the engine through strategic modes, adaptive briefs, tension sliders, and visible tradeoff output. This is the fastest way to feel where the system is strong, thin, useful, or ornamental.
Albert Capability Manual
Current system-level reference for what is operational, what is merely scaffold, what has been validated, and where the real command surface exists today.
Capabilities
16
Integrated across strategy, CRM, messaging, web, memory, monitoring, and recovery
Corpus books
363
Books ingested and chunked across political strategy, economics, history, and theory
Embedded chunks
177,000+
Local semantic index — retrieval-ready. Pipeline running (186k total)
AZHA submissions
Stage 1+2
Live classifier: IHRA definitional gate + 7-dimension social/moral theory scoring
  • Corpus grown to 363 books / 186k chunks. Embedding pipeline running continuously.
  • Decision engine Stage 1 (problem spec gate) + Stage 2 (retrieval + synthesis) fully operational.
  • AZHA antisemitism classifier built: GPT-5.4 Stage 1 (IHRA), Claude Sonnet Stage 2 (theory).
  • Fortnightly corpus acquisition review cron running. Gold-standard calibration set seeded.

Executive map

The system is best understood as perception, judgment, and execution linked by persistent memory and operational tooling.

Perception

  • User requests
  • Political corpus
  • CRM records
  • Memory and wiki notes
  • Dashboard and service state

Judgment

  • Problem structuring
  • Local retrieval and source-pack building
  • Sonnet synthesis
  • Tradeoff and recommendation logic
  • Evaluation harnesses

Execution

  • Reports and PDFs
  • Emails and broadcasts
  • CRM schema and workflows
  • Website deployment
  • Backups, dashboards, and recovery

Capability inventory

Political Strategy Intelligence Assistant

Three-stage engine: Stage 1 problem spec gate (local Qwen + clarifying questions), Stage 2 corpus retrieval (177k+ embedded chunks, max 2 per book for diversity), Stage 3 Sonnet synthesis (diagnosis, 3 options, recommended path, 30-day plan, watchpoints).

Strength
363-book corpus, quality-filtered retrieval, case persistence, and structured output across all runs.
Current limit
Embedding pipeline still completing; fortnightly review cron auditing corpus gaps.

Political Strategy Source Accumulation

363 books ingested (epub, pdf, mobi via Calibre conversion), 186k chunks, 177k+ embedded. Fortnightly acquisition review cron tests 4 problems × 3 perspectives and emails gap report.

Strength
Bulk ingest pipeline, auto-sync from OneDrive, and corpus depth analysis all operational.
Current limit
Anna's Archive automation blocked by DDoS-Guard; manual book drops via OneDrive.

Attio CRM

API-first operational CRM for political, donor, advocacy, and event structures.

Strength
Schema is live and tailored to the Australian political context.
Current limit
Views, pipelines, and live data import still need further work.

Canary Marker + Email

Leak-attributable broadcasts and programmatic email handling are operational.

Strength
Tested end-to-end with encrypted maps.
Current limit
Operational/legal use still requires judgment.

Websites + Intake Infrastructure

Cloudflare Pages and Workers provide autonomous website and form deployment.

Strength
Operational deployment pipeline with analytics and KV-backed forms.
Current limit
Domain purchase still has a manual CAPTCHA step.

Memory + Knowledge

Persistent local semantic memory, wiki compilation, and long-term continuity.

Strength
Zero API-cost memory with semantic recall.
Current limit
Quality still depends on disciplined capture.

Observability + Recovery

Grafana, Prometheus, Loki, watchdogs, and backup protocols now run with much better discipline.

Strength
Current-work state and cost semantics are far more truthful than before.
Current limit
More domain-specific panels can still be added.

AZHA Antisemitism Classifier

Two-stage pipeline for Royal Commission submission analysis. Stage 1: IHRA definitional gate (GPT-5.4 primary, Sonnet fallback). Stage 2: 7-dimension social/moral theory scoring — identity economics, moral utility, moral disengagement, dehumanisation, norm cascades, signalling theory.

Strength
Gold-standard calibration set seeded. Reports generated per-submission and in aggregate.
Current limit
4 gold-standard examples so far; needs ~100 for paper-quality accuracy.

Payments + Donation Flow

Stripe architecture exists but still isn't fully deployed live end-to-end.

Strength
Products, links, and CRM-side model are configured.
Current limit
Webhook worker still not live.

Print & Physical Output

Print-ready PDFs, decks, and reports can be generated for real-world use.

Strength
Operational report/deck rendering paths exist.
Current limit
Printer-device control is still not explicitly integrated/documented.