01 / Opportunity ledger

Opportunities

Defensible gaps ranked by evidence, unexploredness, feasibility, value, and the experiment that could kill them.

Promoted0
Testing0
Watching2
Open / total2 / 5

02 / Ranked ledger

Current candidates

5 records

01verify paper

Provenance-preserving long-horizon memory regression harness

Cross-stack tests that verify memory consolidation and retrieval cannot amplify low-authority observations into high-authority tool actions.

Artificial intelligenceAgent securityDeveloper tools
evidence3unexploredness1technical4operational3value1
02occupied

Generic verbal-reflection middleware

A generic package for reflection, retry, and episodic verbal memory loops.

Artificial intelligenceDeveloper tools
evidence3unexploredness0technical5operational4value2
03occupied

API-Bank compatibility wrapper

A thin commercial wrapper around the released API-Bank tool-use evaluator.

Artificial intelligenceAgent evaluation
evidence3unexploredness0technical5operational4value1
04verify paper

Schema-aware agent safety preflight and regression evidence

Cross-framework regression testing that compares schema-formatted execution with flattened-description safety judgment before tools run.

AI safetyDeveloper toolsSoftware systems
evidence3unexploredness2technical4operational3value2
05occupied

Generic ReAct orchestration product

Commercialize interleaved reasoning-and-acting orchestration as a generic agent architecture.

Artificial intelligenceSoftware systems
evidence3unexploredness0technical5operational4value3