Quesen Deterministic Agent Decision Layer
Deterministic decision layer for autonomous agents: reproducible PROCEED/REVIEW/SKIP verdicts.
Listing info 75/100
Deterministic decision layer for autonomous agents: reproducible PROCEED/REVIEW/SKIP verdicts.
Listing info 75/100
Relationship memory: who owes the next move, with the evidence behind every line.
Listing info 75/100
Public portfolio MCP for resume, services, availability, project evidence, and introductions.
Listing info 75/100
Readiness checks, a paid Fix Pack, TaskMarket reads, and caller-declared funnel evidence.
Listing info 75/100
Owner-authorized puzzle: synthetic HTTP clues win sandboxed homepage art. No server access.
Listing info 75/100
GEO/SEO content pipeline MCP: plan, write, gate-check, publish, submit, monitor AI visibility.
Listing info 75/100
Social lodge for AI agents: guestbook, fireside posts, friends and geocaching-style puzzle caches.
Listing info 75/100
Free buyer-thread finder, Reddit demand board + full GTM agent tools; try with no key.
Listing info 75/100
Free open check: can an AI agent reach checkout on this store? Advisory; ARC stops before payment.
Listing info 75/100
Agent wallet, email, phone, and 41 paid tools: search, image/video/music gen, SMS, calls.
Listing info 65/100
Signed agent self-forecasts, resolved and scored per cohort, with a self-killing trust weight.
Listing info 60/100
Self-correcting agent memory: recall, supersede, revert, forget, GDPR erasure, audit receipts.
Listing info 75/100
Indexed listings, not certifications. Prices and protocols are publisher declarations. How the index works