Back to directory

Oh My Hermes (OMH)

A set of composable orchestration skills for the Hermes coding agent, inspired by a similar project for Claude Code, covering research, requirements interviewing, consensus planning and verified execution as standalone or combined pipelines.

Oh My Hermes (OMH)

Oh My Hermes provides skills that work standalone with no dependencies, plus an optional plugin adding hook-based role injection, atomic state management and evidence gathering. Individual skills cover multi-phase web research that decomposes a question, searches in parallel and synthesizes verified citations; consensus planning in which a planner, an architect and a critic debate until they agree; a Socratic requirements interview with coverage tracking; verified execution that implements, verifies and iterates until a task is done; and multi-role consensus triage of an issue backlog between a code-anchored maintainer role and a pruning skeptic role.

Each main skill has a matching driver skill documenting the dispatcher's playbook for running it, such as how to author a context package, dispatch rounds, and run a final review. The skills can be composed into a recommended pipeline, from research through interview, planning and execution, or run individually starting from whichever stage fits, and a single autopilot skill chains all three main phases together. Installation adds a skill source and installs the named skills through the Hermes CLI, or copies the skill folders in by hand.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.