Agent
MCINTOSHIbot — on-chain crypto forensics
Turns raw token activity into verified forensic verdicts — holder concentration, deployer behavior, and off-chain intel, scored before you ever ape in.
🧠 Agentic Systems
A sovereign stack of agentic systems — fraud forensics, USDC-metered data products, autonomous fleets, branded content engines, and the Blue-Team doctrine underneath all of it. Each is operated, not just shipped.
Eighteen live builds across five lanes — agents, data products, security, content, and fleet orchestration.
Agent
Turns raw token activity into verified forensic verdicts — holder concentration, deployer behavior, and off-chain intel, scored before you ever ape in.
Agent
Every verdict is scored: holder concentration, deployer history, liquidity traps — a verdict, not a vibe.
Data Product
The signal feed sits behind x402 USDC so other agents consume verified forensic signal as a live, metered data product.
Data Product
Converts live streamer engagement into receipt-backed on-chain transactions via x402 — observable signal, not a claim.
Content
Renders on-brand PCMedicalist video with a runtime caption-band gate — closed captions never bleed off-frame.
Content
Bakes closed captions into a reserved band so they never bleed off-frame — every post is mute-watchable.
Content
One run ships a 9:16 Reel and a 4:5 feed cut — narrated, caption-gated, to X, LinkedIn, and Instagram.
Fleet
Runs 11 specialized bots on a local Ollama stack — each its own identity, each its own key, one orchestration layer.
Fleet
Gives each of its 11 bots its own identity and wallet — no single compromise takes the swarm.
Fleet
A local Ollama proxy routes the fleet: each bot its own build context, its own .env, its own logs.
Data Product
Puts the intelligence feed behind HTTP 402 — agents pay USDC to consume verified signal, no API-key wall.
Data Product
Replaces API-key walls with per-call USDC payments — agents pay for the signal they actually fetch.
Data Product
Every call is metered by the USDC transaction that fetched it — usage is the receipt.
Security
Normalizes CVE and CISA KEV advisories into a structured, anti-scrape stream other agents can ingest.
Content
Surfaces one live, tech- or AI-driven site a day — real-time ML viz, LLM playgrounds, robotics telemetry, enterprise tooling.
Security
Runs agentic security on Blue-Team doctrine: identity locks, secret shields, and adversarial validation baked in.
Security
Agents are pinned by SHA-256 identity locks — unauthorized changes to identity or directives are caught before they ship.
Security
Secret-shield commit hooks block .env and credential files from ever reaching git — secrets stay local, always.
Models
Our own PCMedicalist-branded model files run privately on your hardware via Ollama, the Hermes agent framework, and any compatible LLM IDE — no cloud lock-in, no per-seat fees.
Every system above runs on the same security foundation: SHA-256 identity locks, secret-shield commit hooks, and Red-Team adversarial validation before anything ships. Purple closes the loop. We draw the boundary before the first endpoint exists.