DEEP Connects Bold Ideas to Real World Change and build a better future together.
Coming Soon
OmegaClaw GoalChainer connects OmegaClaw deontic directives with PeTTaChainer contextual evidence. It ranks candidate actions against individual and collective goals, flags forbidden actions, and produces a repeatable demo for the BGI Sprint I OmegaClaw improvement track.
OmegaClaw GoalChainer is a goal-aware agent decision layer for the BGI Sprint I OmegaClaw improvement track.
The project connects existing OmegaClaw deontic work with PeTTaChainer contextual evidence. It represents individual and collective goals as weighted requirements, resolves obligations, permissions, prohibitions, and norm conflicts, and returns a ranked action list with proof-oriented metadata.
The current demo is an incident-response planning scenario. Publishing a raw log helps collective coordination, but violates an individual privacy goal and a deontic prohibition. Publishing a redacted summary satisfies privacy, repair, and coordination goals, so the scorer recommends it.
Prototype repo: https://github.com/MesTTo/OmegaClaw-GoalChainer
Related repos: https://github.com/MesTTo/OmegaClaw-Core, https://github.com/MesTTo/omegaclaw-deontic, https://github.com/MesTTo/PeTTaChainer
The full GoalChainer decision layer on PeTTa: norms (lib_deontic), belief (PeTTaChainer), Subjective-Logic (SNARS), and individual-vs-collective motivation (MetaMo) combined into a ranked, proof-backe
The deontic (lib_deontic) and directive (lib_directive) layers, plus the Codex provider that the GoalChainer skill and the demo agent loop run on.
The defeasible + Standard Deontic Logic engine on PeTTa (forbidden / obligated / permitted, with deadlines and contrary-to-duty), with its own test suite.
The PLN contextual query (a truth value with a proof) that grades how acceptable each candidate action is, on the same PeTTa runtime.
© 2026 Deep Initiatives Inc. All rights reserved.
Join the Discussion (0)
Please create account or login to post comments.