Skip to content

Add hermes-atropos-environments bundled skill#815

Merged
teknium1 merged 1 commit into
mainfrom
hermes/hermes-5ab2a29e
Mar 10, 2026
Merged

Add hermes-atropos-environments bundled skill#815
teknium1 merged 1 commit into
mainfrom
hermes/hermes-5ab2a29e

Conversation

@teknium1

Copy link
Copy Markdown
Contributor

Summary

Adds the hermes-atropos-environments skill to the bundled skills directory so it ships with the repo and auto-syncs to users via skills_sync.

What's Included

  • SKILL.md — Comprehensive guide for building RL environments that integrate with Atropos:

    • Architecture overview (HermesAgentBaseEnv → your env)
    • All 6 required methods with code examples
    • Config class patterns
    • Three CLI modes (serve/process/evaluate)
    • 11 common pitfalls
    • Reward function patterns (LLM judge, binary verification, multi-signal)
    • Minimum implementation checklist
  • New: Inference Setup section — Instructs the agent to always ask the user for their inference provider before running any tests:

    1. OpenRouter (+ which model)
    2. Self-hosted VLLM endpoint
    3. Other OpenAI-compatible API
    4. Local Atropos training server

    Includes a provider flags reference table.

  • references/agentresult-fields.md — AgentResult dataclass field reference + common extraction patterns

  • references/atropos-base-env.md — Atropos BaseEnv abstract/overridable methods reference

  • references/usage-patterns.md — Step-by-step testing patterns (process, evaluate, serve, smoke test)

Privacy Review

All content is pure instructional documentation — no API keys, no private URLs, no personal information.

Add comprehensive skill for building, testing, and debugging Hermes Agent
RL environments for Atropos training. Includes:

- SKILL.md: Full guide covering HermesAgentBaseEnv interface, required
  methods, config class, CLI modes (serve/process/evaluate), reward
  function patterns, common pitfalls, and minimum implementation checklist
- New 'Inference Setup' section: instructs the agent to always ask the
  user for their inference provider (OpenRouter + model choice, self-hosted
  VLLM endpoint, or other OpenAI-compatible API) before running tests
- references/agentresult-fields.md: AgentResult dataclass field reference
- references/atropos-base-env.md: Atropos BaseEnv API reference
- references/usage-patterns.md: Step-by-step patterns for process,
  evaluate, serve, and smoke test modes

Will be auto-synced to ~/.hermes/skills/ via skills_sync.
@teknium1 teknium1 merged commit 9be8d88 into main Mar 10, 2026
1 check failed
angelburgosrosado pushed a commit to angelburgosrosado/hermes-agent that referenced this pull request Apr 27, 2026
…ab2a29e

Add hermes-atropos-environments bundled skill
02356abc pushed a commit to 02356abc/hermes-agent that referenced this pull request May 14, 2026
…ab2a29e

Add hermes-atropos-environments bundled skill
olympus-terminal pushed a commit to olympus-terminal/hermes-agent that referenced this pull request May 16, 2026
…ab2a29e

Add hermes-atropos-environments bundled skill
Egavasyug pushed a commit to Egavasyug/hermes-agent that referenced this pull request Jun 10, 2026
…ab2a29e

Add hermes-atropos-environments bundled skill
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant