
Start or resume a work session with full context loading and work tracking. Use at the beginning of work.
On-demand HA-voice brainstorm — reads entity inventory, automation/script listings, and operator intent to surface at most 2 capability-gap ideas, each gated by proposal-triage before becoming a PROP. Invoke when the operator asks "what automations am I missing?", "any coverage gaps?", or "brainstorm improvements". Never runs autonomously.
On-demand dev-voice brainstorm — reads codebase friction signals (git churn, test signal, manifest drift, README coverage) and emits at most 2 improvement ideas, each gated by proposal-triage before becoming a PROP. Invoke when the operator asks "what should I be fixing?", "anything wrong with X?", or "brainstorm improvements". Never runs autonomously.
On-demand fitness-voice brainstorm — reads Strava history and training-goal signals to surface at most 2 coverage-gap or imbalance ideas, each gated by proposal-triage before becoming a PROP. Invoke when the operator asks "what am I neglecting?", "anything off with my training?", or "brainstorm training gaps". Never runs autonomously.
On-demand hermit-voice brainstorm — synthesizes memory, available capabilities, recent compiled artifacts, and codebase shape into at most 2 capability ideas, each gated by proposal-triage before becoming a PROP. Invoke when the operator explicitly asks the hermit to brainstorm capabilities or ideate, e.g. "brainstorm capabilities", "what could you be doing for me?", "any capability ideas?". Never runs autonomously.
Evolves hermit configuration and templates after a plugin update. Detects version gaps, presents new features, walks through new settings. Run after updating the plugin.
Executes the heartbeat checklist from HEARTBEAT.md. Reads the checklist, evaluates each item, and reports findings or acknowledges with HEARTBEAT_OK. Supports run/start/stop/status/edit subcommands.
Lists all proposals with their status, source, and age. Shows auto-detected proposals prominently. Activates on messages like "what have you noticed", "any improvements", "any proposals", "show proposals".
Browse and explain the hermit's Home Assistant automations — list by topic, filter by keyword with plain-language YAML explanations, or sort by last-fired. Read-only. Use when the operator asks "what automations do I have / what does this one do / which haven't fired."
Initializes or resumes a work session. Loads context from OPERATOR.md and SHELL.md, orients the agent, and establishes what to work on. Use at the beginning of every work session.
Fetch and normalize the full Home Assistant state into durable artifacts. Use before drafting automations or when context is stale.
Presence history & tracker-health report — current home/away state, reliability, recent arrival/departure transitions, and activity patterns for person/device_tracker entities. Use when the operator asks about presence history or when a presence-dependent automation (locks, alarm, vacuum, climate) misbehaves.
Evening house brief — end-of-day security check, device status, and energy snapshot. Runs as a daily routine at 22:30 or on demand.
File a GitHub issue or post a comment on an existing issue via a GitHub App bot identity. Use when the operator says "file as a GH issue", "open an issue for PROP-NNN", "report this to the tracker", "add a comment to issue
One-time Home Assistant setup for this hermit. Configures HA access, connects to the official Home Assistant MCP Server integration, and verifies both the Python CLI and HA MCP. Run once per project after /claude-code-hermit:hatch.
Interval scheduled check — detects multi-session cardiac-drift trends across recent compiled activity notes. Runs weekly via the scheduled-checks routine; findings are routed through the proposal pipeline as Evidence Source: scheduled-check/weekly-coaching-patterns.
Captures RPE (1-10) and subjective notes when the operator replies to a strava-sync notification on any configured channel. Triggers on inbound channel messages matching RPE grammar while state/strava-pending-rpe.json has a sync from the last 24h. Channel-agnostic.
Creates a proposal for a high-leverage improvement discovered during work. Only for ideas with real impact — not trivial fixes. Use when you discover something worth operationalizing.
Reflect on recent work and propose improvements if patterns are noticed.
Closes the current work session with a structured handoff. Archives the session report and prepares for the next session. Activates on messages like "I'm done", "wrap it up", "that's it for now", "done for today", "close the session".
Detect dropped HA integrations by computing per-domain unavailable-entity ratios from the latest context snapshot. Flags domains where most entities are unavailable, suggesting the integration lost its connection. Runs daily as a scheduled check via reflect-scheduled-checks.
Operator-sanctioned alternative to bare `git push` from agent context. Open a PR from the current feature branch with a body assembled inline from commit history, the test command's last output, screenshots, and an optional project PR template. Refuses on protected branches, dirty trees, or zero commits ahead. Run as the final step of a ticket.
Accept, defer, dismiss, or resolve a proposal. For accepted proposals, asks how to proceed: start implementing now, create a session task, or note for manual implementation. Activates on messages like "accept PROP-", "dismiss PROP-", "defer PROP-", "resolve PROP-".
Returns a 5-line executive summary of recent work. Checks active session first, falls back to latest report. Activates on messages like "brief", "what happened", "morning update", "overnight summary".
Manages scheduled routines as per-session CronCreate jobs. Each enabled routine in config.json becomes an idle-gated CronCreate registered at session launch.
Per-activity coaching analysis from Strava. Detects interval vs steady-state sessions and road vs trail terrain, branching the metrics accordingly. Computes zone breakdown and recovery estimate; saves a compiled artifact and returns a compact summary. Run after a workout or to retro-analyze a specific activity.
Pre-wrap quality gate. Runs /claude-code-hermit:simplify for a cleanup pass on the working-tree diff (including untracked files), re-runs commands.test, and reports results. Suggests /code-review for a deeper review. Run this before committing.
Initializes the autonomous agent in the current project. Creates the state directory, templates, OPERATOR.md, and config.json. Appends session discipline to CLAUDE.md. Detects installed hermits. Run once per project, like git init.
Activate the dev hermit in the current project. Appends the dev safety rules to CLAUDE.md, installs the git-push-guard hook at strict profile by default, and offers companion plugins. Run once per project after /claude-code-hermit:hatch; re-run to update individual settings.
One-time fitness hermit setup. Configures Strava MCP access, drops routine prompt templates, and wires routines into config.json. Run once per project after /claude-code-hermit:hatch.
Handles inbound messages from Claude Code Channels (Telegram, Discord, webhooks) with session context awareness.
Generate the weekly review report for the current ISO week. Writes to .claude-code-hermit/compiled/review-weekly-YYYY-Www.md and sends a channel-friendly summary with an evolution block. Runs every Sunday at 23:00 via routine.
Parallel code review and cleanup of recent changes. Replicates the original bundled /simplify command that was removed in CC v2.1.146.
Returns an eleven-check health report on the hermit installation — config validity, hook registration, state file integrity, cost budget, proposal health, sibling dependency ranges, file permissions, docker-security overlay drift, archival health, reflect loop health, sandbox capability. Use when diagnosing an install, before a release, or after suspicious behavior. Activates on messages like "/hermit-doctor", "health check", "diagnose the hermit", "what's wrong", "run diagnostic".
Post-hatch validation — checks config, OPERATOR.md, plugin references, routines, and optionally sends a channel test message. Run after hatch to verify setup.
Daily routine that closes the current session at midnight when the operator has been idle ≥10 minutes. Queues a pending close (drained by heartbeat) when the operator is currently active. Intended to fire via the `daily-auto-close` routine; safe to invoke manually.
Audit a hermit-backed repo for safe migration to another machine. Classifies files, generates a migration manifest, and produces a verification checklist. Git-first, conservative by default.
Show alert state, proposal queue depth, routine engagement, and channel availability. Activates on messages like 'health check', 'how's the hermit', 'is anything broken', 'hermit health', 'system health', 'anything wrong', 'hermit status', 'are routines running'.
Generates Docker scaffolding and walks the operator through the full deployment — token setup, build, start, MCP plugin configuration, workspace trust, and verification. Offers to back up and overwrite existing Docker files. Run after /hatch.
Opt-in advanced wizard for Docker security hardening beyond v1.0.26 baseline. Adds LAN containment with DNS policy (firewall + DNS sidecar), resource bounds with kernel hygiene sysctls, and a boot-time plugin install audit log. Each toggle is opt-in with honest cost/benefit framing, applied as a docker-compose overlay (does not modify the base compose file), verified against the live container, and fully reversible. Run after /docker-setup; requires bridge networking.
View or change hermit configuration for this project. Manages model, channels, budget prompts, morning brief, heartbeat, routines, idle behavior, compaction thresholds, Docker packages, and unattended mode.
Show fragile zones, stale accepted proposals, and recent learnings drawn from session history, proposals, and reflect output. Activates on messages like 'what's stuck', 'any fragile zones', 'show me what's blocked', 'recent learnings', 'hermit brain', 'what have you learned lately', 'where are the weak spots'.
Manually record RPE and subjective notes for a specific Strava activity. Use for backfilling, correcting entries, or rating a non-latest activity that the auto-capture skill skipped.
Returns a compact summary of the current session state. Channel-optimized (under 10 lines). Pass --full to append infrastructure health (proposals, routines, last activity, knowledge). Activates on messages like "status", "progress", "what are you working on", "how's it going".
Morning house brief — live status, overnight anomalies, energy snapshot, pending proposals, and today's priorities. Runs as a daily routine or on demand.
Show cost trend, autonomy delta, and proposal-resolution times across recent weeks. Activates on messages like 'how am I trending', 'cost trend', 'autonomy', 'hermit evolution', 'show me my trajectory', 'am I improving', 'proposal velocity', 'weekly trends'.
Analyze Home Assistant history data and entity patterns to identify automation opportunities, unused devices, and energy anomalies. Use periodically or when looking for optimization opportunities.
Check Home Assistant connectivity, context freshness, and locale. Auto-refresh context if stale. Use at session start or when HA status is unclear.
Audit all live Home Assistant automations and scripts against the safety policy. Catches policy drift from entities added via the HA UI that bypassed this plugin's safety gate. Runs weekly as a scheduled check via reflect-scheduled-checks.
Validate and apply a generated HA automation or script YAML with safety checks and optional reload. Use after building or modifying an automation.
Get a quick live status of the house via MCP. Shows lights, covers, climate, presence, and notable states. Use when the operator asks about house state.
Draft a Home Assistant automation or script YAML from a description. Validates against the entity inventory and safety policy. Use when the user wants to create or modify HA automations.
Guided channel activation for local/tmux users — installs the plugin, configures the bot token in the project-local state dir, and walks through pairing. Run after hatch or hermit-settings to activate a configured channel.
Run the configured test suite and record the result to .claude-code-hermit/state/last-test.json. /dev-pr reads this record; a fresh pass at HEAD skips re-running. Use for mid-task verification, debugging a failing test in isolation, or as a building block for /dev-quality.
Delete an automation or script from Home Assistant via REST API with operator confirmation. Use when the operator asks to remove, delete, or disable a specific automation or script.
Check that this project is set up correctly for dev-hermit and safe for the implementer agent. Composes core validators for shared checks; adds dev-specific property verification. Run manually to diagnose setup issues; registered as a weekly scheduled_check (findings-only mode, no side effects).
Lists and offers to clean up stale or merged git branches. Use when the local repo has accumulated feature branches across sessions.
Scan the Home Assistant error log for recurring automation failures. Surfaces any automation that appears in error-flagged log lines three or more times, which typically indicates a silently broken automation. Runs daily as a scheduled check via reflect-scheduled-checks.
Summarizes what the operator did during takeover (via git log), optionally queues instructions in NEXT-TASK.md, updates SHELL.md, and restarts the Docker container. Run locally after a /hermit-takeover session.
Stops the Docker container running the autonomous hermit, marks the session as operator takeover, loads full hermit context, and presents a summary. Run locally (outside Docker) when you want to drive interactively.
Create a feature branch from a clean tree before delegating to the implementer or starting direct edits. Validates working state, picks the right base from protected_branches, and refuses on collisions.
Generate a Monitor entry that tails the project's dev log file for error patterns. Handles date-rotating logs (Winston, Pino, structlog) with a midnight wrapper, and fixed-path logs (Rails log/development.log) with plain tail. Reads dev_log_path_pattern, dev_error_pattern, dev_noise_pattern from config. Run once per project during onboarding.
Knowledge base maintenance — lint raw/ and compiled/ for stale, unreferenced, missing-type, and oversized artifacts. Read-only report with actionable advice. Activates on messages like "check knowledge", "lint knowledge", "knowledge health".
Run the hermit plugin test suites and report pass/fail summary
Background watching via the CC Monitor tool. Starts subprocesses that stream events as conversation notifications — zero token cost when quiet. Supports declared config watches (auto-registered on session start) and ad-hoc operator-invoked watches.
Regenerate obsidian/Connections.md from current hermit state. Runs nightly via routine and after each turn that mutates sessions/proposals. Safe to invoke manually. Cortex Portal.md is live Dataview — no rebuild needed.
Set up the Hermit Cortex — an Obsidian vault surface over your hermit's state. Creates the obsidian/ directory with Brain, Cortex, Evolution, System Health, Connections, and Cortex Portal pages. Run once per project.
Discover this project's test commands, protected branches, dev-server setup (start command, ports, health URL, log path), and stack details. Persists findings under claude-code-dev-hermit.* in config.json and writes a compiled/dev-profile artifact. Run after /hatch or whenever the project's testing or CI setup changes.
One-shot read of branch state, dev-server monitor health, and worktree refs. Read-only — no gates, no fixes; surfaces issues with the recovery command the operator should run. Run any time you want to know "is the server up? what branch am I on? is anything stale?" without scanning three places.
Boot a session-scoped dev server via the core Monitor tool. Reads commands.dev_start, dev_required_ports, dev_health_url, dev_auth_check, and dev_expected_listeners from .claude-code-hermit/config.json. Run before browser-testing or any task that exercises the running app.
Enrich existing hermit content with frontmatter and tags. Scans sessions, proposals, and artifact paths for missing fields, clusters similar files for batch confirmation, then rebuilds Connections.md if the Cortex is set up.
Standalone routine skill for interval-triggered scheduled checks. Runs one due check, gates findings through reflection-judge + proposal-triage, applies state, and logs. Run as a daily routine — not called by reflect.
Stop the session-scoped dev server registered by /dev-up. Runs commands.dev_stop if configured (e.g., docker compose down); otherwise stops the Monitor entry directly. Verifies ports return to free-or-allowlisted state.
Generates Docker scaffolding and walks the operator through the full deployment — token setup, build, start, MCP plugin configuration, workspace trust, and verification. Offers to back up and overwrite existing Docker files. Run after /hatch.
Evolves hermit configuration and templates after a plugin update. Detects version gaps, presents new features, walks through new settings. Run after updating the plugin.
Returns a six-check health report on the hermit installation — config validity, hook registration, state file integrity, cost budget, proposal health, file permissions. Use when diagnosing an install, before a release, or after suspicious behavior. Activates on messages like "/hermit-doctor", "health check", "diagnose the hermit", "what's wrong", "run diagnostic".
Executes the heartbeat checklist from HEARTBEAT.md. Reads the checklist, evaluates each item, and reports findings or acknowledges with HEARTBEAT_OK. Supports run/start/stop/status/edit subcommands.
Handles inbound messages from Claude Code Channels (Telegram, Discord, webhooks) with session context awareness.
Closes the current work session with a structured handoff. Archives the session report and prepares for the next session. Activates on messages like "I'm done", "wrap it up", "that's it for now", "done for today", "close the session".
Generate the weekly review report for the current ISO week. Writes to .claude-code-hermit/compiled/review-weekly-YYYY-Www.md and updates obsidian/Latest Review.md if the cortex is set up. Runs every Sunday at 23:00 via routine.
Reflect on recent work and propose improvements if patterns are noticed.
Audit a hermit-backed repo for safe migration to another machine. Classifies files, generates a migration manifest, and produces a verification checklist. Git-first, conservative by default. In project scope mode, detects unprotected credentials and proposes gitignore additions.
Creates a proposal for a high-leverage improvement discovered during work. Only for ideas with real impact — not trivial fixes. Use when you discover something worth operationalizing.
Initializes the autonomous agent in the current project. Creates the state directory, templates, OPERATOR.md, and config.json. Appends session discipline to CLAUDE.md. Detects installed hermits. Run once per project, like git init.
Returns a 5-line executive summary of recent work. Checks active session first, falls back to latest report. Activates on messages like "brief", "what happened", "morning update", "overnight summary".
Start or resume a work session with full context loading and work tracking. Use at the beginning of work.
Standalone routine skill for interval-triggered scheduled checks. Runs one due check, gates findings through reflection-judge + proposal-triage, applies state, and logs. Run as a daily routine — not called by reflect.
Post-hatch validation — checks config, OPERATOR.md, plugin references, routines, and optionally sends a channel test message. Run after hatch to verify setup.
View or change hermit configuration for this project. Manages model, channels, budget prompts, morning brief, heartbeat, routines, idle behavior, compaction thresholds, Docker packages, and unattended mode.
Initializes or resumes a work session. Loads context from OPERATOR.md and SHELL.md, orients the agent, and establishes what to work on. Use at the beginning of every work session.
Accept, defer, dismiss, or resolve a proposal. For accepted proposals, asks how to proceed — create a session task or note for manual implementation. Activates on messages like "accept PROP-", "dismiss PROP-", "defer PROP-", "resolve PROP-".
Regenerate obsidian/Connections.md from current hermit state. Runs nightly via routine and after each turn that mutates sessions/proposals. Safe to invoke manually. Cortex Portal.md is live Dataview — no rebuild needed.
Enrich existing hermit content with frontmatter and tags. Scans sessions, proposals, and artifact paths for missing fields, clusters similar files for batch confirmation, then rebuilds Connections.md if the Cortex is set up.
Set up the Hermit Cortex — an Obsidian vault surface over your hermit's state. Creates the obsidian/ directory with Brain, Cortex, Evolution, System Health, Connections, and Cortex Portal pages. Run once per project.
Manages scheduled routines as per-session CronCreate jobs. Each enabled routine in config.json becomes an idle-gated CronCreate registered at session launch.
Returns a compact summary of the current session state. Channel-optimized (under 10 lines). Pass --full to append infrastructure health (proposals, routines, last activity, knowledge). Activates on messages like "status", "progress", "what are you working on", "how's it going".
Background watching via the CC Monitor tool. Starts subprocesses that stream events as conversation notifications — zero token cost when quiet. Supports declared config watches (auto-registered on session start) and ad-hoc operator-invoked watches.
Lists all proposals with their status, source, and age. Shows auto-detected proposals prominently. Activates on messages like "what have you noticed", "any improvements", "any proposals", "show proposals".
Run the hermit plugin test suites and report pass/fail summary
Stops the Docker container running the autonomous hermit, marks the session as operator takeover, loads full hermit context, and presents a summary. Run locally (outside Docker) when you want to drive interactively.
Run all plugin test suites, report pass/fail summary
Knowledge base maintenance — lint raw/ and compiled/ for stale, unreferenced, missing-type, and oversized artifacts. Read-only report with actionable advice. Activates on messages like "check knowledge", "lint knowledge", "knowledge health".
Summarizes what the operator did during takeover (via git log), optionally queues instructions in NEXT-TASK.md, updates SHELL.md, and restarts the Docker container. Run locally after a /hermit-takeover session.
Bump the plugin version, write a detailed changelog entry for the upgrade skill to consume, and commit+push. Use this skill whenever the user says "release", "version bump", "cut a release", "changelog and push", or finishes a set of changes and wants to ship them. Also trigger when the user says "do the release thing" or asks to prepare changes for hermits to pick up.
Guided channel activation for local/tmux users — installs the plugin, configures the bot token in the project-local state dir, and walks through pairing. Run after hatch or hermit-settings to activate a configured channel.