Hermes Security
Freshness check
Before writing or changing any Hermes config, run hermes --version and
hermes update --check, then read
https://github.com/NousResearch/hermes-agent/releases/latest. If the installed
or latest version is newer than 0.23.0, read every release note since 0.23.0
("Upgrade notes" first) and adjust. Never apply guidance from this skill to a
newer version without that check.
Native vs pattern
| Need | Status in 0.23.0 | Where |
|---|---|---|
| Human approval of dangerous terminal commands | Native (approvals.mode) |
approval-model.md |
| Never-run list that survives YOLO | Native (approvals.deny + hardline blocklist) |
approval-model.md |
| Per-call approval of an MCP tool | Native (trust: untrusted, spend_limit) |
approval-model.md |
| Staged memory / skill writes | Native, opt-in (*.write_approval) |
approval-model.md |
| Isolation of agent commands | Native backends (terminal.backend) |
sandbox-backends.md |
| Secrets out of the model's reach | Native (.env, secret sources, redaction, env sanitizing) |
secrets.md |
| No email/message leaves without the owner's go | Pattern - not a Hermes setting | locked-send-pattern.md |
| Private vs company sealing | Partly native (profiles) + Pattern (filesystem, credentials) | privacy-partition.md |
Critical rules
- Never tell an owner "Hermes will ask before sending". Sending
(
hermes sendfrom the terminal, cron delivery, the Email adapter's replies) is not an approval-gated action. Enforce it with the go gate (locked-send-pattern.md). (send_messageis not registered as a model tool at v0.23.0: tools/send_message_tool.py L28-30, toolsets.py L186-187.) - A rule the model can break is a wish. Enforcement lives in OS
permissions, missing credentials,
approvals.denyand sandbox mounts - SOUL.md rules only explain the intent. - Never
approvals.mode: off/ YOLO on a profile that acts for an owner.cron_mode/single_query_mode: approveis set on the chief (root profile) only; manager and worker profiles keep the defaultdeny- there is no inheritance from the root config (approval-model.md). - Container backends skip dangerous-command approval - the container is the
boundary, so audit its mounts,
docker_forward_envand network. - Write guards are not a sandbox: write_file/patch are guarded, the terminal is not. The agent's OS user must not be able to read any send credential.
terminal.env_passthroughreaches both the terminal and Hermes cron scripts - a send credential never goes there.- Review
command_allowlistafter every setup session; each "Always" widens access on every surface, cron included. - Never ship or install a send script. After the go, the send runs through the mechanism already on the owner's server; audit it (sends only after an explicit go? from which account?) and report. No mechanism → an owner decision in the checklist, never an invented one.
Workflow
- Freshness check;
hermes security audit;hermes doctor. - Merge security-config.yaml
into each agent profile; adjust
approvals.denyto the CLIs installed. - Pick the backend per profile (sandbox-backends.md).
- Move secrets per secrets.md;
chmod 600each.env. - Apply the go gate and audit the existing send mechanism (locked-send-pattern.md); run its verification steps.
- Seal profiles (privacy-partition.md) and run its cross-read test.
Sources
- https://hermes-agent.nousresearch.com/docs/user-guide/security
- https://hermes-agent.nousresearch.com/docs/user-guide/configuration (Smart Approvals, Security, Terminal Backend Configuration)
- https://hermes-agent.nousresearch.com/docs/reference/mcp-config-reference
- https://hermes-agent.nousresearch.com/docs/user-guide/secrets/
- https://hermes-agent.nousresearch.com/docs/user-guide/multi-profile-gateways (per-profile isolation)
- https://hermes-agent.nousresearch.com/docs/user-guide/features/cron (script credentials)
- https://github.com/NousResearch/hermes-agent/releases/tag/v0.23.0