Dashboard

Your AI-fluency control room. Learn the map, scan the field, ship upgrades.

What's new

freshest techniques from your watchlist

Next up

Fresh upgrades

Learn

Six tracks, high altitude. Read → recall → apply. Mark each lesson as you go.

Review

Active recall with spaced repetition. Grade honestly — that's what makes it work.

Methodology

X Intel Scanner

Automated: pick a source, set what to track, and it scans on a timer — no pasting. Manual capture is also available below.

…
no scan yet
Heuristic classifier (offline). Set AIK_LLM=1 to enrich via the claude CLI.

Corpus

Review & Push

Only proven techniques (credible source + corroboration + grounded + low hype) reach this queue, auto-attributed to the projects they fit. Read the caveats, then confirm to push a ready brief into each project (.claude/ai-upgrades/). Credibility ≠ proof — confirm real value by testing (eval-gate playbook).

Playbooks

Curated, project-agnostic techniques — drop-in skills, CLAUDE.md rules, and hooks that improve how any project runs. Install a skill or rule straight into one of your projects, or copy a hook snippet.

Suggested for your projects · auto, from what's trending per project

Briefs

Project-ready markdown auto-generated from your most actionable scanned content. Filter by project, then copy or download to drop straight into a project's CLAUDE.md or a skill.

Sources & Mapping

Add the accounts and topics you want tracked, and control which of your projects each kind of technique maps to. Changes apply to the next scan.

Add a source

X account → nitter feed (mirror-rotated). Topic → Hacker News search feed (works cloud-side). Raw URL → used as-is.

Active sources

Project mapping

For each technique (concept), tick the projects it should generate upgrades for. Untick all = applies to every project.

Project upgrades

Concrete changes mined from intel, mapped to your Claude projects. Apply or dismiss each.