330da12b5e
All checks were successful
CI / Shell unit tests (push) Successful in 1m33s
Secret scan / Scan diff for credential-shaped strings (pull_request) Successful in 16s
CI / Shell unit tests (pull_request) Successful in 1m45s
CI / validate (pull_request) Successful in 8m58s
CI / validate (push) Successful in 9m11s
The github.com personal account HongmingWang-Rabbit was suspended on 2026-05-06 along with the Molecule-AI org, so the previous Dockerfile URL is 404. Image rebuilds have been broken since. Migrates the install source to the new Gitea-canonical home: https://git.moleculesai.app/molecule-ai/hermes-agent Both branches (main, feat/platform-adapter-plugins) and 9 release tags were pushed from the local clone preserving full history. The HERMES_FORK_REF default (feat/platform-adapter-plugins) is unchanged and resolves to the same SHA (18e4849e) as before. Aligns with feedback_per_agent_gitea_identity_default and reference_post_suspension_pipeline. Tracked in molecule-ai/internal#72. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|---|---|---|
| .github/workflows | ||
| .molecule-ci/scripts | ||
| docs | ||
| runbooks | ||
| scripts | ||
| tests | ||
| __init__.py | ||
| .gitattributes | ||
| .gitignore | ||
| .runtime-version | ||
| adapter.py | ||
| CLAUDE.md | ||
| config.yaml | ||
| Dockerfile | ||
| executor.py | ||
| install.sh | ||
| known-issues.md | ||
| README.md | ||
| requirements.txt | ||
| start.sh | ||
template-hermes
Molecule AI workspace template that runs the real Nous Research hermes-agent behind an A2A bridge.
What's actually in the workspace
Both the Docker path and the SaaS bare-host path run the same stack:
- hermes-agent — real upstream project, installed via
scripts/install.shfrom NousResearch/hermes-agent. Gateway boots with the OpenAI-compatible API server platform enabled on127.0.0.1:8642(internal only). - molecule_runtime — A2A server + bridge adapter. Listens on
:8000and forwards every incoming message to the local hermes-agent gateway. Canvas, plugins, skills installer see the same A2A contract as any other runtime.
Two execution paths
This template ships two entrypoints because the platform has two execution models — see internal/product/designs/workspace-backends.md for the full story.
| Path | Used by | Entry | Install recipe |
|---|---|---|---|
| Docker (1 container / workspace) | docker compose up local dev |
ENTRYPOINT ["start.sh"] in Dockerfile |
Image build: RUN curl install.sh | bash in Dockerfile |
| Bare host (1 EC2 / workspace) | SaaS production | /opt/molecule-venv/bin/molecule-runtime (CP user-data) |
install.sh runs at workspace-provision time as ubuntu user |
start.sh and install.sh do the same logical work (install
hermes-agent, seed ~/.hermes/.env + config.yaml, start hermes gateway, wait for :8642). They stay symmetric; when you change
one, check the other.
This template was rewritten in v2.0.0 — the previous version was a thin
OpenAI-compat provider shim that shared the hermes name with the real
project but had none of its agent capabilities (skills, memory, tools,
self-improvement loop, multi-platform gateway). See
docs/PLANNING.md for the full rewrite
rationale.
Usage
In Molecule AI canvas
Select this template when creating a new workspace — the canvas
Runtime dropdown resolves hermes to workspace-template:hermes
via molecule-monorepo/workspace-server/internal/provisioner/provisioner.go.
From a URL (community install)
github://Molecule-AI/template-hermes
Required environment
At least one provider key must be set, matching whichever model you select in the Config tab. hermes-agent picks the right one by prefix — you do not pick the provider yourself.
| Env var | Used for |
|---|---|
HERMES_API_KEY |
Nous Portal (Hermes 3/4 direct) |
OPENROUTER_API_KEY |
Anything via OpenRouter (200+ models) |
ANTHROPIC_API_KEY |
Claude direct (native SDK inside hermes-agent) |
OPENAI_API_KEY |
GPT direct |
GEMINI_API_KEY |
Gemini direct (native SDK inside hermes-agent) |
MINIMAX_API_KEY |
MiniMax direct |
Set these as workspace-level secrets (POST /settings/secrets) — see
molecule-monorepo/docs/runbooks/saas-secrets.md for the canonical
flow.
Persisting skills and memory
hermes-agent writes to ~/.hermes (/home/agent/.hermes in the
container). Mount this path as a persistent volume if you want skills,
memory, and cron schedules to survive workspace restarts — the
platform's default Docker named volume does this automatically as long
as the workspace isn't re-provisioned from scratch.
Files
| File | Purpose |
|---|---|
Dockerfile |
Docker-path: builds the image (hermes-agent + molecule_runtime) |
start.sh |
Docker-path entrypoint: boots hermes gateway, waits for :8642, exec's runtime |
install.sh |
Bare-host path (EC2/SaaS): runs at provision time as the runtime user. Installs hermes-agent + starts gateway in background. Called by CP user-data after pip-install of molecule_runtime, before molecule-runtime launches. |
adapter.py |
HermesAgentAdapter(BaseAdapter) — just a factory |
executor.py |
HermesAgentProxyExecutor — A2A → hermes HTTP bridge |
config.yaml |
Template metadata + model list for the Config tab |
requirements.txt |
Python deps for the bridge (molecule_runtime + httpx) |
docs/PLANNING.md |
Rewrite plan + rationale + phase breakdown |
docs/ARCHITECTURE.md |
How the bridge works, port map, failure modes |
docs/MIGRATION.md |
Upgrade path from v1.x (the old adapter shim) |
docs/CONFIGURATION.md |
How to pick a model, rotate keys, tune hermes-agent |
Schema version
template_schema_version: 1 — compatible with Molecule AI platform v1.x.
License
Business Source License 1.1 — © Molecule AI. hermes-agent itself is
MIT-licensed by Nous Research and installed from its upstream repo at
build time.