oh-my-hermes is a skill for Hermes Agent that transforms natural-language operator requests into explicit goals, constraints, and plans, while enforcing evidence boundaries between prepared intent and verified results. It also separates project-memory candidates from approved records so only reviewed context is recalled.
Project overview
By enforcing deterministic local routing, catalogs, manifests, and claim rules, oh-my-hermes addresses the problem of AI agents presenting prepared intent as verified execution, which is a structural integrity concern for agent-driven automation workflows.
Project type
AI Agent · AI Coding
Use cases
Coding & Development
Deployment
Refer to project documentation
License
MIT
Best for
Developers and AI engineers who need AI agents to produce explicit goals, structured plans, and honest execution status from natural-language requests
Key capabilities
Turns ambiguous requests into explicit goals, constraints, tradeoffs, acceptance criteria, and a plan that can be handed off.
Separates prepared intent, observed runtime events, and verified results to prevent claiming execution occurred without proof.
Keeps project-memory candidates separate from approved records and recalls only reviewed context.
Limitations and risks
The skill does not read or mutate opaque Hermes memory; it only manages project-memory candidates separately from approved records.
A prepared brief is never presented as a generated or visually verified artifact.
Getting started
Run curl -fsSL https://raw.githubusercontent.com/rlaope/oh-my-hermes/main/install.sh | sh, then run omh setup. Setup difficulty is rated easy due to the single-command install script or skill tap path.