Explore Project fit
Pick a repo question. The map returns the public skill, project route, proof boundary, and visible gap an agent can inspect.
Search by problem, capability, or evidence question. The answer is a source trail, not a sales claim.
try:
▲ technical◆ hybrid● creativedashed ring = private work, public overview only
Explore Public project index
graph click → target card → source
Can a dream give a persona more than memory can?
Preregistered study: dreams as inspectable intermediates whose certified affect drives the persona’s live voice — emotion tags, tone, identity held stable.
check script present
How does an agent prove which browser tab actually answered?
Browser transport for agents — authenticated Chrome, element refs, screenshots, proven tab provenance.
check script present
What do you do with exploit code you can't trust? Breed it.
Red/blue exploit evolution in isolated Docker targets, scored by deterministic proof gates.
check script present
How do you build an agent harness that doesn't trust agents?
Memory-first zero-trust harness: every agent turn gets a goal, evidence, receipt, and a stop condition.
check script present
What if extraction had no knobs to mislead you with?
Zero-choice document extraction: one file in, one truthful result envelope out.
check script present
What did the research sweep fail to reach?
Multi-source research with traceable citations and honest used/skipped/degraded provider reporting.
check script present
What was on screen at that exact moment?
Video into timestamped frames, transcript, and reports — evidence agents can re-query from memory.
check script present
How many times should provider glue be written? Once.
One execution surface for any model — chat, image, batch, tools, agents, receipts.
check script present
What was true at the moment the bug happened?
Real breakpoints and paused-frame state — prove runtime truth before patching.
check script present
What should an agent remember before it touches the code?
BM25 + Qdrant candidates + ArangoDB graph recall for agents.
check script presentprivate evidence
Who decides, when the AI sounds sure? Humans do.
Space-cyber evidence threads traced to human decisions — relevance never becomes authority.
external repoprivate evidence