hermitd
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Check it deeper on your machine
It has changed since its scan on 17 Nov 2026. Its scripts were only partly scanned.
As a Claude Code plugin, install it and then ask Claude Code:
/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/gtapps/hermitd with scanskillsafety
Or from a terminal, clone it and run it:
git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/gtapps/hermitd'
It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.
Badge for your README
Markdown
[](/listings/gtapps--hermitd)HTML
<a href="/listings/gtapps--hermitd"><img src="/api/listings/gtapps--hermitd/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
These steps run a script from the network that findsafeskills does not scan.
**Choose one installation method below.** Run it from the folder where you want your agent, empty or existing. Uses your Claude subscription on Linux, macOS, or Windows via WSL2. See [prerequisites](plugins/hermitd/docs/how-to-use.md#prerequisites).
<details open> <summary>Already run Claude Code? Install the plugin</summary>
With Claude Code 2.1.325+ and Bun 1.4+ installed:
claude plugin install hermitd --marketplace gtapps/hermitd --scope local claude "/hermitd:hatch"
</details>
<details> <summary>Fresh machine? Use the bootstrap installer</summary>
Prepares Claude Code, Bun, and tmux, installs the plugin, and launches setup:
curl -fsSL https://gtapps.github.io/hermitd/install.sh | bash
</details>
Both options set up the agent in this folder. Hatch guides you through the agent’s purpose and preferences, then shows how to start it. Choose Quick for defaults you can adjust later.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
.claude-plugin/marketplace.json · Manifest · 157 lines
- {
- "name": "hermitd",
- "owner": {
- "name": "gtapps"
- },
- "metadata": {
- "description": "Your own local Claude Tag, in the Claude app, Discord, Telegram, or a custom Claude Code channel."
- },
- "plugins": [
- {
- "name": "hermitd",
- "source": "./plugins/hermitd",
- "description": "Your own local Claude Tag, in the Claude app, Discord, Telegram, or a custom Claude Code channel.",
- "version": "1.4.34",
- "category": "productivity",
- "author": {
- "name": "gtapps",
- "url": "https://github.com/gtapps"
- },
- "license": "MIT",
- "homepage": "https://github.com/gtapps/hermitd",
- "repository": "https://github.com/gtapps/hermitd",
- "keywords": [
- "always-on",
- "assistant",
- "automation",
- "claude-tag",
- "gtapps"
- ]
- },
- {
- "name": "hermitd-dev",
- "source": "./plugins/hermitd-dev",
- "description": "Git safeguards, debugging, and merge-conflict resolution for your Hermit.",
- "version": "0.4.27",
- "category": "development",
- "author": {
- "name": "gtapps",
- "url": "https://github.com/gtapps"
- },
- "license": "MIT",
- "homepage": "https://github.com/gtapps/hermitd/tree/main/plugins/hermitd-dev",
- "repository": "https://github.com/gtapps/hermitd",
- "keywords": [
- "development",
- "git-safety",
- "code-review",
- "gtapps"
- ]
- },
- {
- "name": "hermitd-homeassistant",
- "source": "./plugins/hermitd-homeassistant",
- "description": "Build, manage, monitor, and troubleshoot your Home Assistant setup with Hermit.",
- "version": "0.4.28",
- "category": "home-automation",
- "author": {
- "name": "gtapps",
- "url": "https://github.com/gtapps"
- },
- "license": "MIT",
- "homepage": "https://github.com/gtapps/hermitd/tree/main/plugins/hermitd-homeassistant",
- "repository": "https://github.com/gtapps/hermitd",
- "keywords": [
- "home-assistant",
- "automation",
- "mcp",
- "gtapps"
- ]
- },
- {
- "name": "hermitd-fitness",
- "source": "./plugins/hermitd-fitness",
- "description": "Fitness coaching for Hermit, with read-only Strava access, activity analysis, training plans, and recovery flags.",
- "version": "0.2.10",
- "category": "fitness",
- "author": {
- "name": "gtapps",
- "url": "https://github.com/gtapps"
- },
- "license": "MIT",
- "homepage": "https://github.com/gtapps/hermitd/tree/main/plugins/hermitd-fitness",
- "repository": "https://github.com/gtapps/hermitd",
- "keywords": [
- "fitness",
- "strava",
- "training",
- "mcp",
- "gtapps"
- ]
- },
- {
- "name": "hermitd-feed",
- "source": "./plugins/hermitd-feed",
- "description": "Scheduled briefs for Hermit from sources you choose, with story tracking, weekly summaries, and source-quality checks.",
- "version": "0.2.8",
- "category": "productivity",
- "author": {
- "name": "gtapps",
- "url": "https://github.com/gtapps"
- },
- "license": "MIT",
- "homepage": "https://github.com/gtapps/hermitd/tree/main/plugins/hermitd-feed",
- "repository": "https://github.com/gtapps/hermitd",
- "keywords": [
- "feed",
- "brief",
- "rss",
- "trending",
- "gtapps"
- ]
- },
- {
- "name": "hermitd-laravel-forge",
- "source": "./plugins/hermitd-laravel-forge",
- "description": "Manage Laravel Forge servers, sites, deployments, and logs with Hermit, with approval before changes.",
- "version": "0.0.24",
- "category": "development",
- "author": {
- "name": "gtapps",
- "url": "https://github.com/gtapps"
- },
- "license": "MIT",
- "homepage": "https://github.com/gtapps/hermitd/tree/main/plugins/hermitd-laravel-forge",
- "repository": "https://github.com/gtapps/hermitd",
- "keywords": [
- "laravel",
- "forge",
- "deployment",
- "devops",
- "php",
- "gtapps"
- ]
- },
- {
- "name": "hermitd-scribe",
- "source": "./plugins/hermitd-scribe",
- "description": "A GitHub App Bot for Hermit that turns proposals and reports into issues, with privacy checks and approval before publishing.",
- "version": "0.1.12",
- "category": "maintainer",
- "author": {
- "name": "gtapps",
- "url": "https://github.com/gtapps"
- },
- "license": "MIT",
- "homepage": "https://github.com/gtapps/hermitd/tree/main/plugins/hermitd-scribe",
- "repository": "https://github.com/gtapps/hermitd",
- "keywords": [
- "github",
- "issues",
- "bot",
- "maintainer",
- "gtapps"
- ]
- }
- ]
- }
README.md · README · 361 lines · 1 finding marked
- <p align="center">
- <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-blue.svg" alt="MIT License" /></a>
- <a href="https://code.claude.com/docs/en/plugins"><img src="https://img.shields.io/badge/Claude%20Code-plugin-orange.svg" alt="Claude Code Plugin" /></a>
- <a href="plugins/hermitd/CHANGELOG.md"><img src="https://img.shields.io/badge/version-1.4.34-green.svg" alt="Version 1.4.34" /></a>
- <img src="https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/gtapps/hermitd/_gh_traffic_stats/.github/badges/clones.json" alt="Downloads" />
- <img src="https://img.shields.io/badge/PRs-welcome-brightgreen.svg" alt="PRs Welcome" />
- <a href="https://discord.gg/54sJqAxhUh"><img src="https://img.shields.io/badge/Discord-Join-5865F2?logo=discord&logoColor=white" alt="Join" /></a>
- </p>
- # Like Claude Tag, but local.
- Run an **always-on Claude Code agent** on your own machine, for **you** or **your team**.
- **Multi-project orchestration**: The agent can start and orchestrate sessions from any projects it can reach. Take over the same session in the Claude app or the CLI. Tasks, saved watches, and pending decisions survive a restart.
- Connect **Discord, Telegram, iMessage, or a custom Claude Code [channel](https://code.claude.com/docs/en/channels)**. Use the Claude app via [Remote Control](https://code.claude.com/docs/en/remote-control), or your terminal.
- The agent has its **own browser ([agent-browser](https://github.com/vercel-labs/agent-browser))**. Take over and Hand back when you want, for example a login. Or use [Chrome on your laptop](https://code.claude.com/docs/en/chrome), including when the agent runs on another machine.
- Give it ongoing responsibilities: maintain research, monitor systems, run routines, and follow up on unfinished work. Between requests, it checks those responsibilities, carries progress across sessions, and reaches you when something needs attention.
- Run it on your **Claude subscription** and extend it with your own MCP servers, skills, and plugins.
- **It's Claude Code.** hermitd is a plugin for Claude Code, no new harness, no wrapper around Claude Code. You get all of Claude Code.
- https://github.com/user-attachments/assets/84012a4d-4ca1-475e-b82d-1db45634a5e3
- ## How it compares
- | Option | Why hermitd |
- |--------|-------------|
- | [Hermes Agent](https://github.com/NousResearch/hermes-agent), [OpenClaw](https://github.com/openclaw/openclaw) | hermitd runs directly inside Claude Code as a plugin, without a wrapper or separate agent loop. You get Claude.ai, Artifacts, the Claude app, [Remote Control](https://code.claude.com/docs/en/remote-control), and the rest of Claude Code, with your tools, permissions, skills, plugins, and MCP. |
- | [Claude Projects](https://code.claude.com/docs/en/claude-projects.md) | **Multi-project orchestration**, with your CLIs and tools. Chat through any Claude Code [channel](https://code.claude.com/docs/en/channels). **Significant cheaper** on routines: [Monitor](https://code.claude.com/docs/en/tools-reference#monitor-tool) and a [precheck](plugins/hermitd/docs/routine-authoring.md) can skip a run without model tokens. |
- | [Claude Tag](https://claude.com/docs/claude-tag/concepts/how-it-works) | Use any **Claude Code channels: Discord, Telegram, iMessage, or a custom channel**, and continue the same session in the Claude app or terminal. |
- | [Claude Duet](https://support.claude.com/en/articles/17631495-get-started-with-claude-duet), [ChatGPT Dots](https://learn.chatgpt.com/docs/dots), [Grok Bot](https://docs.x.ai/grok-bot/overview) | It runs locally, **fully customizable**. Routines can be **significant cheaper**. |
- | [Self-hosted Claude Managed Agents](https://claude.com/resources/articles/self-hosted-claude-managed-agents) | hermitd runs on **your Claude subscription** rather than API usage, which can be 10x cheaper. |
- <a id="quick-start"></a>
- ## Set up
- **Choose one installation method below.** Run it from the folder where you want your agent, empty or existing. Uses your Claude subscription on Linux, macOS, or Windows via WSL2. See [prerequisites](plugins/hermitd/docs/how-to-use.md#prerequisites).
- <details open>
- <summary>Already run Claude Code? Install the plugin</summary>
- With Claude Code 2.1.325+ and Bun 1.4+ installed:
- ```bash
- claude plugin install hermitd --marketplace gtapps/hermitd --scope local
- claude "/hermitd:hatch"
- ```
- </details>
- <details>
- <summary>Fresh machine? Use the bootstrap installer</summary>
- Prepares Claude Code, Bun, and tmux, installs the plugin, and launches setup:
- ```bash
- curl -fsSL https://gtapps.github.io/hermitd/install.sh | bash
- Note: Pipes a downloaded script directly into a shell (curl | sh / wget | bash)
- Line 63 on GitHub
- ```
- </details>
- Both options set up the agent in this folder. Hatch guides you through the agent’s purpose and preferences, then shows how to start it. Choose Quick for defaults you can adjust later.
- ## Keep it running
- After setup, follow the printed next steps to start your agent.
- ### On your machine
- Run in a persistent tmux session:
- ```bash
- hermitd start
- ```
- Requires tmux. The watchdog recovers failed sessions while your machine stays on. Claude Code's `/sandbox` is recommended for unattended use. To connect a chat, run `/hermitd:channel-setup` as directed by the setup handoff.
- [Host setup and operations](plugins/hermitd/docs/always-on-ops.md)
- ### In Docker
- Run the guided setup in Claude Code:
- ```text
- /hermitd:docker-setup
- ```
- Builds and starts the container, then walks you through authentication and channel pairing. Requires Docker Compose v2.
- [Docker setup](plugins/hermitd/docs/always-on.md)
- **Customize the container.** Ask the agent to add tools, packages, or services to its Docker setup. For example: “Add ffmpeg to the container.”
- Optional [Docker security controls](plugins/hermitd/docs/docker-security.md) cover local-network access, DNS policy, resource limits, and plugin installation auditing.
- ## What the plugin adds
- - **Continuity.** Persistent working state and archived session handoffs carry progress across compaction and restarts. An external watchdog recovers failed sessions, while context management keeps long-running sessions manageable.
- - **Proactive work.** Heartbeats regularly check the responsibilities you give the agent. Routines run scheduled work, and watches surface changes. Together, they let the agent follow up without waiting for another request.
- - **Work through chat.** Assign work and receive results in your connected chat. Longer assignments get threaded progress updates, with a separate reply when the agent needs a decision. Assignments can also carry a persistent task record with requester, due date, result confirmation, and a dashboard view by person.
- - **Your network.** Local databases, SSH, LAN devices, filesystem are reachable because the agent runs where they are.
- - **Token efficiency.** With Claude Code’s [Monitor](https://code.claude.com/docs/en/tools-reference#monitor-tool), heartbeat checks and optional routine prechecks run outside the model. Quiet checks and skipped routines use no model tokens; eligible routines due together can share a wake.
- - **Lasting knowledge.** Turn source material in `raw/` into maintained knowledge in `compiled/`, alongside Claude Code's auto memory. `/recall` searches past sessions, knowledge, proposals, and captured channel conversations.
- - **Learning from experience.** The agent reviews evidence from its work and operation, saves useful lessons, and verifies proposed behavior changes before bringing them to you for approval.
- - **Control and visibility.** Track progress, pending decisions, and usage through the dashboard. Pause is enforced at the tool boundary, and optional usage caps can alert you or pause further work.
- **Part of your project channel.** With passive mode, the agent saves incoming group messages to look back on later, and wakes when someone you allow @mentions it. It also remembers instructions for that channel. Fixed policies you set for a chat change only through an approval prompt, and an optional scheduled review proposes rule updates when the same correction keeps coming up.
- **Work that belongs to a chat.** Routines, watches, and reminders created from a chat report back to that chat. A team chat lists and edits its own routines, and changing one it does not own asks for approval. [Third-party channel plugins](plugins/hermitd/docs/channel-plugins.md) can join in: a small send script gives them proactive notices and routine chats, and Docker approves them for you.
- <a id="configure-it"></a>
- ## Configure
- Tune from a terminal with `/hermit-settings`, or change permitted settings from a trusted connected chat. Every write is validated and recorded in a redacted audit ledger; `/hermit-settings history [setting]` shows what changed. Some of the settings available:
- | Key | Default / options (default **bold**) |
- |-----|--------------------------------------|
- | `agent_name` | your assistant's name |
- | `operator_profile` | primary-chat audience: **`technical`** / `non-technical` |
- | `timezone` | machine's zone stored at setup; must be usable by the scheduler |
- | `language` | unset until chosen or learned in the first conversation |
- | `escalation` | how much it does before asking: `conservative` / **`balanced`** / `autonomous` |
- | `model` | session model: **`sonnet`** |
- | `effort` | session reasoning effort: **unset** (Claude Code default) |
- | `permission_mode` | how freely the unattended agent acts: **`auto`** |
- | `AGENT_HOOK_PROFILE` | guardrail profile: `minimal` / **`standard`** (interactive) / **`strict`** (always-on) |
- | `channels` | Discord / Telegram / iMessage / third-party channel plugins (+ `allowed_users`) |
- | `channels.primary` | which channel gets outbound pings |
- | `channels.<name>.maintainer_channel_id` | optional separate chat for technical alerts, diagnostics, and usage details |
- | `push_notifications` | native/mobile push on alerts: **`true`** |
- | `remote` | remote control; `false` also requires approval for cross-machine peer messages; **`true`** |
- | `approvals` | auto-mode approval policy: **`personal`** (existing behavior) or `team` (typed turns and DM-paired approvers in the maintainer chat; chat prompts close after 60 seconds); effective at next start |
- | `ask_gate` | route unattended questions to a paired channel: **`true`** |
- | `budget` | optional daily / weekly / monthly caps; **`alert`** or binding `pause` action |
- | `artifacts` | dashboard / weekly review: **dashboard enabled** |
- | `heartbeat.enabled` | timed idle sweeps: **`true`** |
- | `heartbeat.every` | idle sweep cadence: **`30m`** |
- | `heartbeat.active_hours` | active window: **`08:00`–`23:00`** |
- | `heartbeat.model` / `heartbeat.effort` | heartbeat evaluator: **`haiku`** at **`high`** |
- | `routines` | persistent routines managed via `/hermit-routines` |
- | `monitors` | persistent background watches managed via `/watch` |
- | `scheduled_checks` | session-triggered skills at task completion |
- | `reflection.graduation_min_sessions` | proposal recurrence bar: **`1`** |
- | `quality_gate.tier` | post-change cleanup spend: **`budget`** / `balanced` / `quality` |
- | `knowledge.compiled_budget_chars` | fresh/resumed startup catalog budget: **`2500`** |
- | `knowledge.raw_retention_days` | `raw/` retention: **`14`** |
- | `knowledge.working_set_warn` | warn above N compiled docs: **`20`** |
- | `auto_session` | auto-start session on boot: **`true`** |
- | `boot_skill` / `shutdown_skill` | custom boot / teardown skill |
- | `context_hygiene.clear` | safe-boundary context clear: enabled, quiet **`1h`**, max age **`24h`**, minimum **20,000** tokens |
- | `context_hygiene.compact` | compact long-running active context: **enabled**, `100000` compactible tokens / `4h` cooldown |
- | `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE` | auto-compact at % of context: **`65`** |
- | `MAX_THINKING_TOKENS` | thinking-token cap per turn: **`10000`** |
- | `CLAUDE_CODE_DISABLE_AGENT_PROGRESS_SUMMARIES` | skips subagent spinner-label requests: **`1`** |
- | `watchdog.scheduler_enabled` | OS scheduler for the watchdog tick: **`true`** on tmux always-on (auto-installed at boot); `false` or `hermitd-watchdog uninstall` opts out |
- | `watchdog.enabled` | recovery/restart tier: **`false`** until first scheduler registration (or `/docker-setup`); hygiene still runs |
- Full schema in the [Config Reference](plugins/hermitd/docs/config-reference.md)
- **Your own prompt.** Replace the agent's built-in prompt and rules with `.hermit/resident-prompt.md`, and spawned helpers' with `.hermit/helper-prompt.md`. Delete the file to return to the shipped prompt. See [Replacing the resident prompt](plugins/hermitd/docs/how-to-use.md#replacing-the-resident-prompt).
- ## Observe
- **Artifacts.** The agent uses [Claude Code Artifacts](https://code.claude.com/docs/en/artifacts) to provide an interactive dashboard and custom pages generated on demand that you can view, interact with, and share. Ask it to build your own personalized agent dashboard.
- Ask for an update from your terminal or connected chat:
- | Command | What it gives you |
- |---------|-------------------|
- | `/brief` | Current status and a summary of recent work. |
- | `/recall` | Search past sessions, knowledge, proposals, and captured conversations. |
- | `/hermit-health` | Alerts, routines, channels, blockers, and recent learnings. |
- | `/hermit-doctor` | Diagnostics for the installation, runtime, scheduling, credentials, and permissions. |
- | `/hermit-evolution` | Cost trends, proposal activity, routines, and what the agent has produced over time. |
- | `/cost-reflect` | A breakdown of usage by token type, session, and what triggered the work. |
- | `/hermit-dashboard-design` | A dashboard designed around what your agent actually tracks. |
- ## Learning loop
- The agent learns along three paths: closed work can save or revise its own learned skills when skill learning is enabled, a nightly dream verifies and tidies memory, and reflection keeps durable lessons and brings other behavior changes to you for approval.
- ```text
- Work produces evidence
- (task records, transcripts, cost logs)
- │
- ┌─────────────┼──────────────┐
- ▼ ▼ ▼
- Task closes Nightly dream Reflect when due
- │ │ │
- ▼ ▼ ▼
- Review the Verify and Verify the cited
- procedure tidy memory evidence
- │ and topic │
- ▼ pages (undo) ┌───┴────┬─────────┐
- Save or patch ▼ ▼ ▼
- a learned skill Remember Note it Candidate
- (pin, rollback) a lesson until it change
- recurs │
- ▼
- Drop duplicates,
- Monthly brainstorm ──► weak cases, and
- what memory covers
- │
- ▼
- You approve?
- │ │
- no yes
- │ │
- No change Re-check,
- then implement
- │
- ▼
- A check passes, or
- you accept the result
- ```
- Reflection runs daily when something is due, after routines configured to reflect, and on scheduled skill checks. Suggestion paths check memory first, so nothing memory already records is proposed again. Close-time learning reviews delegated and longer tasks and edits only learned skills; a needed routine or agent becomes a proposal. Proposals are [task records](plugins/hermitd/docs/task-records.md) waiting for admin consent.
- ## Ask it from your chat
- | Ask | What it does |
- |-----|--------------|
- | “Start X in `~/code/api` on Opus at high effort, in its own worktree.” | Starts a session in any project, on your model and effort. |
- | “Tell me when my session in `~/code/web` finishes.” | Watches your other Claude Code sessions. |
- | `!pause` · `!snooze 2h` · `!model sonnet` · `!effort high` | Claude Code controls, straight from chat. |
- | “`/later` check tomorrow whether those errors have returned.” | Checks whether a fix held up over time. |
- | “remind me Monday at 14:00 to pay the invoice” | Delivers a durable reminder at the due minute in monitor mode, even after a restart. |
- | “Download last month's invoices from the supplier portal.” | Uses its [own managed browser](plugins/hermitd/docs/how-to-use.md#browser-tasks) by default and offers phone Take over and Hand back on Docker and Linux/WSL2 tmux. The operator's Chrome is opt-in for their account. |
- | “When I ask for a status update, include blockers.” | Remembers instructions for that channel. |
- | “What else could you be doing for me?” | Proposes new capabilities from your work and its tools. |
- ## Cost
- Quiet heartbeat checks, skipped routines, and passive chat capture use no model tokens. Work, evaluations, and replies consume usage; context management keeps conversation history bounded.
- ```text
- event: a file, a log line, a webhook, a schedule
- │
- ▼
- precheck (plain script, no model tokens)
- │
- ├── nothing changed ──▶ back to sleep
- │
- ▼ something to do
- Claude Code turn on your machine
- │
- ├── needs a decision ──▶ asks you in chat
- │
- ▼
- result in your chat + usage recorded
- ```
- - **See what drives usage.** Token usage is recorded per call, including the model, input/output/cache split, and whether work came from a routine, heartbeat, channel, or another source. Session and daily totals feed the dashboard, weekly review, and `/cost-reflect`.
- - **Set limits.** Optional daily, weekly, and monthly caps can alert you or enforce a pause until the exceeded budget window resets. Under Claude subscription billing, dollar figures are usage estimates rather than additional per-token charges.
- - **Choose where to spend.** Set the session model and effort, and optionally a model and effort per routine or per task. Heartbeat checks run on Haiku at high effort by default. Routine models run in isolated subagents, so use them for work that can return a concise result.
- See [budgets](plugins/hermitd/docs/config-reference.md#budget) and [routine scheduling](plugins/hermitd/docs/routine-authoring.md) for configuration and scheduler fallback behavior.
- ## Remote work
- Reach the running agent through your connected channels or Claude Code Remote Control. You can also start separate sessions for additional work:
- - **Background sessions with follow-up.** Through [`/spawn-session`](plugins/hermitd/skills/spawn-session/SKILL.md), the agent launches a local Claude Code helper in the project, or in another folder with `--cwd`, and relays its status when it becomes idle. Claude Code isolates the helper's edits in a Git worktree unless the project sets `worktree.bgIsolation` to `none`; pass `--worktree` to give it one from the start. Set the helper’s model and effort with options such as `--model sonnet --effort high`.
- - **Task threads for helpers.** A task thread can be bound to a spawned session: follow-ups in the thread go to that session, its reports land on the task, and `!model` or `!effort` in the thread switch that helper without losing its conversation. A request that continues the same work goes back to the same session.
- - **Local [Remote Control](https://code.claude.com/docs/en/remote-control) gate.** Through [`/rc-gate`](plugins/hermitd/skills/rc-gate/SKILL.md), the agent manages a Remote Control server on your machine or server. While the gate is open, you can spawn new Claude Code sessions from the Claude app, using your local files and tools. Each session gets its own Git worktree, while the agent keeps running.
- Both session-spawning paths require a Git workspace. Remote Control requires a Claude sign-in through `/login` on the machine running the agent.
- **Watch other sessions.** Through [Claude Code cross-session messaging](https://code.claude.com/docs/en/cross-session-messaging), ask the agent to watch a local Claude Code session, including one you started interactively, and notify you when it next becomes idle.
- **Claude Code controls from chat.** Use `!model sonnet`, `!effort high`, `!advisor opus`, `!compact`, `!clear`, `!doctor` (alias `!checkup`), and `!permission-mode auto` directly from your connected chat. Control the agent’s work with `!pause`, `!resume`, and `!snooze 2h`. Use [`/when-done-switch-to --model sonnet`](plugins/hermitd/skills/when-done-switch-to/SKILL.md) to switch automatically at the end of the current turn.
- ## Extensions
- Optional plugins that add domain tools and workflows to your agent.
- - [dev-hermit](plugins/hermitd-dev/README.md): Branch discipline, push guards, and gated PR workflows.
- - [homeassistant-hermit](plugins/hermitd-homeassistant/README.md): Home Assistant tools, automation workflows, and safety checks.
- - [fitness-hermit](plugins/hermitd-fitness/README.md): Strava integration, activity analysis, and training routines.
- - [hermitd-feed](plugins/hermitd-feed/README.md): Source curation, recurring briefs, and weekly synthesis.
- - [hermitd-laravel-forge](plugins/hermitd-laravel-forge/README.md): Laravel Forge deployments, logs, and server management.
- - [hermitd-scribe](plugins/hermitd-scribe/README.md): GitHub issues and comments from proposals through a dedicated bot identity.
- You can run separate agents for different responsibilities, each with its own working state, knowledge, and routines. See [Creating Your Own Hermit](plugins/hermitd/docs/creating-your-own-hermit.md).
- **External orchestration.** Other agents and tools can check the agent’s status, health, and recent work through its [MCP interface](plugins/hermitd/docs/external-control-surface.md), and request a wake when needed.
- ## Maintenance
- **Sign-in renewal from chat.** When your agent’s Claude sign-in needs renewing, use `/relogin` from your connected chat. Open the link in your browser, sign in, and send the code back in chat.
- **Tool sign-ins.** Anyone admitted may ask to reconnect a tool or claude.ai connector; the agent tells an admin, who finishes the sign-in with `/mcp` at the agent's terminal. The account used serves every chat and routine, so prefer a dedicated account over a personal login, and forward the link `/mcp` shows if the tool belongs to someone else's account.
- **Scheduled backups.** Optional backups preserve the agent’s knowledge, session reports, settings, and Claude Code memory in Git, with an optional private remote copy. Backups run without model tokens.
- ## Upgrading
- Run `hermitd update` from the project folder, or `hermitd update <name>` from anywhere. Docker updates refresh the host core first, then the container. The update waits for a quiet moment, applies migrations, then restarts the agent once and resumes the conversation. External updates finish with a `hermitd doctor --fix` pass and record its result in update history. Run `hermitd doctor [--fix]` to list health findings and their commands, or apply safe repairs and re-check; queued upgrades remain pending until the agent applies them.
- `hermitd list` shows registered and discovered agents on this host, including stopped and missing projects. `hermitd status [name]` shows transport, execution and its age, open and waiting tasks, and the task in progress, including one a worker or helper runs. Both support `--json`. Listing never removes entries; `hermitd prune` removes missing projects.
- Use `hermitd start|stop|restart|attach [name]` for lifecycle commands, `hermitd pause [name] on|off|snooze <duration>|status`, `hermitd watchdog [name] run|install|uninstall`, or `hermitd run [name] <script> [args]` for maintenance. Names match the project folder or agent name; with no name, the nearest project above the current folder is used.
- `hermitd agents [name]` opens Claude Code's agent view (`claude agents`) where the agent runs: inside the container for Docker agents, on the host under the agent's Claude config folder otherwise. On the host that view lists every session in the config folder, so pass `--cwd <path>` to narrow it. Extra flags pass through, and `--json` works from scripts.
- `hermitd skills [name] status|pin|unpin|rollback|archive|restore [skill]` inspects or controls learned skills. Use `status --skill <skill>` for one bounded history and its last use. Mutations run inside the container when the Docker agent is up. A task close can save a reusable procedure automatically; "keep this skill" pins it and "stop doing it that way" about a named learned skill rolls back its latest revision.
- See the [Upgrade guide](plugins/hermitd/docs/upgrading.md) for details.
- <a id="tips--tuning"></a>
- ## Guides
- - **Configure:** the [Config Reference](plugins/hermitd/docs/config-reference.md) covers the full schema and tuning details.
- - **Use:** [Getting Started](plugins/hermitd/docs/how-to-use.md) and the [Owner's Guide](plugins/hermitd/docs/owners-guide.md) cover everyday work, decisions, and controls.
- - **Automate:** [Routine Authoring](plugins/hermitd/docs/routine-authoring.md) covers schedules and prechecks. [Channel configuration](plugins/hermitd/docs/config-reference.md#channels) includes third-party channel plugins.
- - **Observe:** [Artifacts](plugins/hermitd/docs/artifacts.md) explains the dashboard, pending decisions, and weekly reviews. [OpenTelemetry](plugins/hermitd/docs/opentelemetry.md) explains how to export each step to a collector you run. Event names and attributes are in [Claude Code's monitoring docs](https://code.claude.com/docs/en/monitoring-usage).
- - **Maintain:** [Upgrading](plugins/hermitd/docs/upgrading.md), [Backup](plugins/hermitd/docs/backup.md), [Troubleshooting](plugins/hermitd/docs/troubleshooting.md), and [Uninstalling](plugins/hermitd/docs/how-to-use.md#install).
- - **Understand:** [Architecture](plugins/hermitd/docs/architecture.md), [Security](plugins/hermitd/docs/security.md), and [FAQ](plugins/hermitd/docs/faq.md).
- [All documentation](plugins/hermitd/docs/)
- ## Community
- Join the [Discord community](https://discord.gg/54sJqAxhUh) for setup help and discussion. See [CONTRIBUTING.md](CONTRIBUTING.md) for reporting bugs or contributing.
- ## Credits
- [Andrej Karpathy](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f) inspired the `raw/` → `compiled/` knowledge system.
- [Building always-on agents](https://claude.com/resources/articles/building-always-on-agents) describes the surrounding architecture: an event inbox, timers, memory, permissions, and an outbox.
- [The components of agent memory](https://www.anthropic.com/research/components-of-agent-memory) describes a small index loaded every session, with the detail behind it.
- [Continual-ARC](https://arxiv.org/abs/2611.14505) measures skill acquisition over a working lifetime for agent harnesses (Lenton, Arif, and Mahmood, 11 Nov 2026). The authors' write-up is [Why we built Continual-ARC](https://unify.ai/blog/why-we-built-continual-arc).
- [Securing persistent agents like OpenClaw at OpenAI](https://developers.openai.com/blog/securing-persistent-agents-like-openclaw-at-openai) describes how OpenAI isolates, sandboxes, and identifies the persistent agents in its own deployment.
- [Hermes Agent](https://github.com/NousResearch/hermes-agent) and [OpenClaw](https://github.com/openclaw/openclaw) are other open source agents.
- [agent-browser](https://github.com/vercel-labs/agent-browser) from Vercel is the managed browser. [Claude in Chrome](https://code.claude.com/docs/en/chrome) is the alternative.
- ## License
- [MIT](LICENSE)
- Independent project, not affiliated with Anthropic.
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.