hermes
Skill2 starsChanged 6 months agoMITView on GitHub
This skill should be used when self-hosting, installing, configuring, operating, updating, troubleshooting, or learning to use NousResearch Hermes Agent, including CLI usage, gateway and Slack setup, plugins, tools, skills, profiles, memory, MCP, cron, Kanban, delegate_task subagents, dashboards, and tmux-managed Hermes sessions.
- Reads credentials
2 other projects use a version of this.
Adopted by
Most recently taken up by
- matheussluzz/ZeuZ-Agenttheir own version, first seen Jul 15, 2026
- aks129/HealthClawGuardrailstheir own version, first seen Jul 4, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- aks129/HealthClawGuardrails★ 33same name
- matheussluzz/ZeuZ-Agent★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 2 are distinctive, and 8 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
8
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 10
Use this skill for Hermes Agent self-hosting and day-to-day operation: install, setup, model/provider selection, gateway platforms, Slack …
- Very specific: line 12
Do not rely on stale remembered commands.
- Distinctive: line 18
If command behavior matters, also refresh the generated local command snapshot:
- Very specific: line 28
Verify local state with scripts/hermes_check.py.
- Very specific: line 29
Use hermes --help and hermes <command> --help as the source of truth for executable commands.
- Distinctive: line 47
Prefer deterministic commands and scripts for inspection.
- Very specific: line 78
Avoid old third-party examples built around hermes run.
- Very specific: line 82
Treat ~/.hermes/.env, Slack tokens, OAuth tokens, and ~/.hermes/auth.json as secrets.
- Very specific: line 84
Use hermes update --check before update decisions.
- Very specific: line 86
Prefer files over chat for durable Hermes behavior: config.yaml, .env, SOUL.md, AGENTS.md, skills, and runbooks.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Hermes
Use First
Use this skill for Hermes Agent self-hosting and day-to-day operation: install, setup, model/provider selection, gateway platforms, Slack channel setup, command discovery, updates, plugins, tools, skills, memory, profiles, MCP, cron, Kanban, delegate_task subagents, logs, dashboards, tutorials, and long-running tmux sessions.
Do not rely on stale remembered commands. Hermes changes quickly. At the start of each Hermes task, resolve scripts/hermes_check.py relative to this skill directory and run:
python3 <skill-dir>/scripts/hermes_check.py --jsonIf command behavior matters, also refresh the generated local command snapshot:
python3 <skill-dir>/scripts/hermes_check.py --write-help <skill-dir>/references/command-map.generated.mdReport update availability before changing configuration. Do not run hermes update without explicit user approval unless the user already asked to update.
Decision Flow
- Verify local state with
scripts/hermes_check.py. - Use
hermes --helpandhermes <command> --helpas the source of truth for executable commands. - Load only the needed reference file:
references/setup-and-update.mdfor install, versioning, updates, config health, backups, and safe update policy.references/getting-started-tutorials.mdfor first-run setup, self-hosting paths, guided usage flows, and user-facing tutorials.references/command-catalog.mdfor command categories and common CLI patterns.references/official-links.mdfor official docs, release, reference, and troubleshooting links.references/slack-gateway.mdfor Slack app, channel, Socket Mode, manifest, gateway, and allowlist setup.references/operations-runbook.mdfor tmux sessions, one-shot execution, profiles, cron, webhooks, memory, skills, and autonomy guardrails.references/kanban-operations.mdfor durable task boards, profile workers, dependencies, isolated workspaces, dispatch, and task notifications.references/delegation-vs-kanban.mdfordelegate_task, subagent delegation, and choosing between delegation, Kanban, tmux, cron, and background terminal work.references/slash-commands.mdfor in-session/...commands, gateway slash commands, and when to prefer CLI commands.references/providers-tools-security.mdfor key paths, config sections, providers, toolsets, approval modes, redaction, and privacy.references/plugins-and-tools.mdfor user/operator plugin installation, tool enablement, MCP connections, and safe verification.references/voice-and-media.mdfor STT, TTS, voice mode, image input, and media-related troubleshooting.references/self-improvement.mdfor updating this skill after new Hermes behavior, command changes, or operational lessons are discovered.references/troubleshooting.mdfor logs, Slack failures, token problems, stale skills, gateway issues, and community-known pitfalls.references/command-map.generated.mdonly when exact local help output is needed.
- For interactive Hermes or
tmuxwork, read the current pane before sending keys. Explain prompts and recommend a choice before sending input. - Prefer deterministic commands and scripts for inspection. Keep non-deterministic guidance in reference docs or in the Hermes conversation itself.
- At the end of each Hermes task, run the self-improvement review in
references/self-improvement.md.
Common Request Routing
- "Install Hermes" or "self-host Hermes" ->
getting-started-tutorials.md,setup-and-update.md, thenofficial-links.md. - "Set up Slack channel" ->
slack-gateway.md, runscripts/hermes_slack_check.py, thentroubleshooting.md. - "Check if my Hermes is healthy" -> run
scripts/hermes_ops_check.py, thensetup-and-update.md. - "Enable a plugin/tool/MCP server" ->
plugins-and-tools.md,providers-tools-security.md. - "Teach me how to use Hermes" ->
getting-started-tutorials.md,slash-commands.md,command-catalog.md. - "Use Kanban", "dispatch workers", or "coordinate profiles" ->
kanban-operations.md,operations-runbook.md. - "Use delegate_task", "delegate this", "spawn subagents", or "parallel subtasks" ->
delegation-vs-kanban.md; use Kanban instead when the work must be durable, human-visible, restartable, or assigned to named profiles.
Current CLI Baseline
As of Hermes Agent v0.12.0, prefer:
hermes
hermes chat -q "single query"
hermes -z "script-friendly one-shot"
hermes setup
hermes model
hermes doctor
hermes config check
hermes update --check
hermes gateway setup
hermes gateway status
hermes slack manifestAvoid old third-party examples built around hermes run; current Hermes does not expose run as a root command in v0.12.0.
Safety Rules
- Treat
~/.hermes/.env, Slack tokens, OAuth tokens, and~/.hermes/auth.jsonas secrets. Never print full token values. - Set or verify allowlists before enabling Slack or other gateways.
- Use
hermes update --checkbefore update decisions; usehermes backupor the update backup flags for risky changes. - For automation, start with read-only/reporting workflows, add verification, then schedule with cron/webhooks only after repeated successful manual runs.
- Prefer files over chat for durable Hermes behavior:
config.yaml,.env,SOUL.md,AGENTS.md, skills, and runbooks.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

