hermit-settings

name	hermit-settings
description	View or change hermit configuration for this project. Manages model, channels, morning brief, heartbeat, routines, idle behavior, compaction thresholds, Docker packages, and unattended mode.
disable-model-invocation	true

Hermit Settings

View or modify the hermit configuration for this project.

Step 0 — Channel reply

If this skill was invoked from a channel-arrived message (the inbound prompt contains a <channel source="..."> tag), reply via that channel's reply tool. Otherwise emit to conversation.

On a channel-tagged turn, every free-form Ask: prompt below is delivered via the reply tool instead of waiting on terminal input — the branch proceeds as an over-channel exchange (ask, then act on the reply when it arrives), the same as any other channel conversation. Never call AskUserQuestion on a channel-tagged turn — it renders in the terminal, invisible to a remote operator. The one bounded ask in this skill (quality-gate, below) additionally queues a durable micro-proposal entry per channel-responder § Channel-safe ask bridge (schema: reflect § Queuing procedure), so it survives compaction or a session restart; free-form asks queue nothing.

Usage

/claude-code-hermit:hermit-settings              — show all current settings
/claude-code-hermit:hermit-settings name          — set agent name
/claude-code-hermit:hermit-settings language       — set preferred language
/claude-code-hermit:hermit-settings timezone       — set timezone
/claude-code-hermit:hermit-settings escalation     — set escalation threshold
/claude-code-hermit:hermit-settings sign-off       — set sign-off line
/claude-code-hermit:hermit-settings channels       — configure channels
/claude-code-hermit:hermit-settings remote          — toggle remote control
/claude-code-hermit:hermit-settings model           — set Claude model
/claude-code-hermit:hermit-settings brief          — configure morning brief
/claude-code-hermit:hermit-settings permissions    — configure unattended mode
/claude-code-hermit:hermit-settings heartbeat      — enable/disable, interval, quiet mode, active hours
/claude-code-hermit:hermit-settings watchdog       — enable/disable, stale_factor, escalate_after, operator_grace, context hygiene compaction
/claude-code-hermit:hermit-settings routines        — manage scheduled routines (add/edit/remove/enable/disable)
/claude-code-hermit:hermit-settings idle             — set idle behavior (wait or discover)
/claude-code-hermit:hermit-settings env              — view/edit environment variables
/claude-code-hermit:hermit-settings compact          — configure SHELL.md compaction thresholds
/claude-code-hermit:hermit-settings docker           — view/edit Docker packages
/claude-code-hermit:hermit-settings scheduled-checks    — manage scheduled plugin skill checks
/claude-code-hermit:hermit-settings boot-skill       — view/clear/change the always-on boot skill
/claude-code-hermit:hermit-settings quality-gate     — set post-implementation /claude-code-hermit:simplify gate tier (budget|balanced|quality)
/claude-code-hermit:hermit-settings reflection       — tune graduation threshold (graduation_min_sessions)
/claude-code-hermit:hermit-settings push-notifications — toggle PushNotification doorbell (fires when no channel is enabled or a configured channel is unreachable)
/claude-code-hermit:hermit-settings artifact-dashboard — toggle the Hermit Dashboard artifact (single-URL status/proposals/weekly-evolution page)
/claude-code-hermit:hermit-settings artifact-proposals — toggle the Proposals-page artifact (full-text open-proposal page with deep-linked anchors)
/claude-code-hermit:hermit-settings artifact-weekly-review — toggle the Weekly-review artifact (stable-URL passthrough of the compiled weekly report)
/claude-code-hermit:hermit-settings artifact-authorization — record the unattended Artifact publish decision (applied by hermit-start at boot, not from this session)

Plan

1. Read config

Read .claude-code-hermit/config.json. If it doesn't exist, inform the operator: "No config found. Run /claude-code-hermit:hatch first."

Scalar and enum edits below are written through scripts/settings-edit.ts, which read-modify-writes the whole config (preserving every sibling key) and refuses a malformed file. Shorthand used in this skill:

bun ${CLAUDE_PLUGIN_ROOT}/scripts/settings-edit.ts .claude-code-hermit/config.json get [dotted.path]      # dump whole config, or one value
bun ${CLAUDE_PLUGIN_ROOT}/scripts/settings-edit.ts .claude-code-hermit/config.json set <dotted.path> <value>   # 'none'/'clear' → null; value is JSON-parsed then falls back to raw string
bun ${CLAUDE_PLUGIN_ROOT}/scripts/settings-edit.ts .claude-code-hermit/config.json toggle <dotted.path>       # boolean flip (absent → true)

2. Show or modify

If no argument (or argument is "all"): Dump the current config with settings-edit ... get (no path), then display all current settings in a readable format:

Hermit Settings (.claude-code-hermit/config.json)

Identity:
  Agent name:      Atlas          → any string | 'none' to clear
  Language:        pt             → any locale code (e.g. en, pt, es, fr)
  Timezone:        Europe/Lisbon  → any tz (e.g. UTC, America/New_York)
  Escalation:      balanced       → balanced | conservative | autonomous
  Sign-off:        Atlas out.     → any string | 'none' to clear

Operational:
  Channels:        none           → run: /claude-code-hermit:hermit-settings channels
  Remote control:  disabled       → yes | no
  Model:           default        → opus | sonnet | haiku | none (for default)
  Morning brief:   disabled       → run: /claude-code-hermit:hermit-settings brief
  Idle behavior:   discover       → discover | wait
  Heartbeat:       disabled       → yes | no  (interval, active hours, stale threshold)
  Watchdog:        disabled       → yes | no  (stale_factor, escalate_after, operator_grace)
  Routines:        2 configured   → run: /claude-code-hermit:hermit-settings routines
  Quality gate:    budget         → budget | balanced | quality
  Permission mode: auto           → default | acceptEdits | auto | plan | dontAsk | bypassPermissions
  Push notif:      disabled       → on | off
  Auto session:    enabled        → read-only
  Boot skill:      /claude-code-hermit:session  → any namespaced skill | 'none' to reset to default
  tmux name:       hermit-myproject → read-only

Compaction:
  Monitoring:      compact at 30 lines, keep 20
  Session Summary: compact at 30 lines, keep 15
  → run: /claude-code-hermit:hermit-settings compact

Environment (env):
  AGENT_HOOK_PROFILE              standard
  CLAUDE_AUTOCOMPACT_PCT_OVERRIDE 65
  MAX_THINKING_TOKENS             10000
  → run: /claude-code-hermit:hermit-settings env

Scheduled Checks:
  automation-recommender  claude-code-setup     interval  7 days   2026-04-01  enabled
  md-audit                claude-md-management  interval  7 days   (never)     enabled
  md-revise               claude-md-management  session   —        2026-04-06  enabled
  → run: /claude-code-hermit:hermit-settings scheduled-checks

Docker:
  Packages: build-essential, ffmpeg  → run: /claude-code-hermit:hermit-settings docker

If argument is "name": Ask: "Agent name? (e.g., Atlas, Hermit, Scout, or 'none' to clear) [current value or skip]" Run settings-edit ... set agent_name <value> ("none"/"clear" → null).

If argument is "language": Auto-detect the system locale via Bash as a default suggestion. Ask: "Preferred language? (e.g., pt, en, es, fr) [current value or auto-detected]" Run settings-edit ... set language <value>. Then re-sync the artifact-chrome translation table (the dashboard/proposals pages overlay .claude-code-hermit/state/artifact-strings.json per key over the English defaults):

• New value is not en: emit bun ${CLAUDE_PLUGIN_ROOT}/scripts/artifact-strings-scaffold.ts <value> <current-ISO-timestamp>, translate every strings value into that language (keep keys and {placeholder} tokens verbatim), and write .claude-code-hermit/state/artifact-strings.json.
• New value is en: delete .claude-code-hermit/state/artifact-strings.json if it exists (absent file ⇒ English chrome).

If argument is "timezone": Auto-detect the system timezone via Bash as a default suggestion. Ask: "Timezone? (e.g., Europe/Lisbon, America/New_York, UTC) [current value or auto-detected]" Run settings-edit ... set timezone <value>.

If argument is "escalation": Ask: "How autonomous should your assistant be?

1. Balanced — act on routine tasks, ask for significant changes (default)
2. Conservative — ask before most non-trivial actions
3. Autonomous — proceed unless blocked, minimize interruptions

Choose 1-3: [current value]" Run settings-edit ... set escalation <conservative|balanced|autonomous>.

If argument is "sign-off": Ask: "Sign-off line for channel messages and briefs? (e.g., 'Atlas out.', '— A.', or 'none' to clear) [current value or skip]" Run settings-edit ... set sign_off <value> ("none"/"clear" → null).

If argument is "channels": Show current channel configuration from config.json → channels object. The channels.primary key (if set) is a magic pointer to the preferred outbound channel, not a channel itself — display it on its own line above the channel list:

Channels:
  Primary: discord    (or "none — falls back to first eligible channel in config order")
  discord  enabled  allowed_users: [123456789]  morning_brief: 07:00  state_dir: /abs/path/...
  (or "No channels configured")

Ask: "Add, remove, edit, or set primary? (add discord / add telegram / remove / edit / primary / primary clear / done) [done]" Loop until operator says "done":

• add : Create entry channels.<name>: { "enabled": true, "dm_channel_id": null }. Prompt for allowed_users (paste user ID or skip) and state_dir (relative or absolute path — defaults to .claude.local/channels/<name>). Set state_dir in the channel entry. Note: "Configure the channel token next: Docker → /claude-code-hermit:docker-setup; tmux or interactive → /claude-code-hermit:channel-setup."
• remove : Delete channels.<name> from config.json. If channels.primary === <name>, also delete channels.primary (a dangling pointer would fail validation) and tell the operator: "Also cleared channels.primary (was pointing at the removed channel)."
• edit : Sub-menu — "What to change? (allowed_users / morning_brief / enabled / done)"
  • allowed_users: "Paste user IDs (space-separated), or 'clear' to allow everyone, or 'block' for empty array." Update channels.<name>.allowed_users.
  • morning_brief: "Enable morning brief for this channel? (yes / no) [current]". If yes: channels.<name>.morning_brief = { "enabled": true, "time": "<HH:MM>" }. If no: set to null.
  • enabled: Toggle channels.<name>.enabled.
• primary : Validate <name> exists as a key in channels (and is not primary itself). If valid, set channels.primary = "<name>". If invalid, reject: "No channel named <name> configured. Add it first with add <name>."
• primary clear: Delete channels.primary. Outbound sends will fall back to the default discord → telegram → imessage order. Note: "Channel changes take effect on next hermit-start run. channels.primary is consulted live by scripts/resolve-outbound-channel.ts on every proactive send — no restart needed for that key alone."

If argument is "remote": Ask: "Enable remote control? Connect from a browser or phone via claude.ai/code. yes — enable remote control no — disable remote control [current: ]" Run settings-edit ... set remote <true|false> (or toggle remote if the operator just wants to flip it). Note: "Remote control changes take effect on next hermit-start run."

If argument is "model": Ask: "Claude model to use? opus → claude-opus-4-6 sonnet → claude-sonnet-4-6 haiku → claude-haiku-4-5-20251001 none → use Claude Code default (inherit from user config) [current: <value or 'default'>]" Run settings-edit ... set model <value>. To clear (operator says "none", "default", or "clear" → inherit Claude Code default), pass the none sentinel: settings-edit ... set model none (the script maps none/clear → null; do NOT pass the literal word default, which would be stored as a string). Note: "Model changes take effect on next hermit-start run."

If argument is "brief":

• If no channels configured: "Morning brief requires channels. Configure channels first with /claude-code-hermit:hermit-settings channels."
• If channels configured:
  • Show current morning_brief setting per channel: channels.<name>.morning_brief
  • Ask: "Enable morning brief delivery? (yes / no) [current value]"
  • If yes: Ask "What time? (e.g., 07:00) [current or 07:00]" and "Which channel? [current or first enabled channel]"
  • Update channels.<selected-channel>.morning_brief: { "enabled": true, "time": "<HH:MM>" } in config.json. If no, set to null.

If argument is "boot-skill": Ask: "Boot skill to invoke on always-on launch? This runs after heartbeat/routines when the tmux session starts. Domain hermits (e.g. claude-code-homeassistant-hermit) declare their own — /claude-code-homeassistant-hermit:ha-boot. Leave as none to use the default (/claude-code-hermit:session). — any namespaced skill (e.g. /claude-code-foo-hermit:foo-boot) none — clear (falls back to /claude-code-hermit:session) [current: <value or 'default'>]" Run settings-edit ... set boot_skill <value>. To clear (operator says "none", "default", or "clear" → fall back to /claude-code-hermit:session), pass the none sentinel: settings-edit ... set boot_skill none (the script maps none/clear → null; do NOT pass the literal word default, which would be stored as a string). The domain boot skill is responsible for calling /claude-code-hermit:session-start itself — this setting just controls the single bootstrap command hermit-start.ts fires into the REPL. Note: "Boot skill changes take effect on next hermit-start run."

If argument is "permissions": Ask: "Permission mode for Claude Code? (auto / acceptEdits / default / plan / dontAsk / bypassPermissions) [current value]"

• auto — autonomous mode; a classifier reviews each action before it runs. Generally available to all users across subscription plans and API usage; supported models and provider configuration can vary. If Claude reports it unavailable for the current selection, choose a supported model or another permission mode. (default)
• acceptEdits — auto-approves file edits, prompts for shell commands
• default — prompts for permission on first use of each tool
• plan — read-only exploration, no file modifications or shell commands
• dontAsk — denies all tools not in permissions.allow; requires a curated allowlist in settings.json
• bypassPermissions — no checks; isolated containers/VMs only
• Note: auto may report "unavailable" at launch if your plan/model/provider doesn't qualify — see Permission Modes Run settings-edit ... set permission_mode <value>.

If argument is "heartbeat":

• Show current heartbeat config (including stale_threshold)
• Ask: "Enable background heartbeat? (yes / no) [current: ]"
• If yes: show the configurable sub-fields before asking each one:

  Heartbeat sub-fields (press Enter to keep current value):
    interval  — how often to check (e.g. 5m, 15m, 30m)         [current]
    active    — active hours window (e.g. 08:00-23:00)          [current]
    stale     — alert if no session progress for (e.g. 2h, 30m) [current]
  

  Then ask each field in sequence.
• Write each changed field through settings-edit ... set heartbeat.<field> <value> (heartbeat.enabled, heartbeat.every, heartbeat.active_hours.start, heartbeat.active_hours.end, heartbeat.stale_threshold). Per-field dotted sets preserve the untouched siblings (waiting_timeout, clean_recheck_cooldown, model).
• After the change is written, reconcile the live Monitor with the new state via the Skill tool. The monitor's poll interval is baked in at start time from heartbeat.every, so a config-only change otherwise leaves the running monitor on the old cadence (and /hermit-doctor would flag the mismatch). Surface the result inline:
  • If heartbeat is now enabled: invoke /claude-code-hermit:heartbeat start (idempotent — stops the old monitor and re-registers at the new interval). Success: "Heartbeat monitor restarted at new interval (<every>). Active immediately."
  • If heartbeat is now disabled: invoke /claude-code-hermit:heartbeat stop. Success: "Heartbeat monitor stopped."
  • Failure (either): "Settings saved to config.json, but /claude-code-hermit:heartbeat <start|stop> failed: . Run it manually to apply."

If argument is "watchdog":

• Show current watchdog config from config.json:

  Watchdog (config.json watchdog)
  
    enabled                false
    stale_factor           2
    escalate_after         3
    operator_grace         15m
    context_clear_tokens   700000
  
  Context hygiene compact (config.json context_hygiene.compact)
  
    enabled                true
    min_context_tokens     150000
    min_interval           4h
  

• Ask: "Enable watchdog? (yes / no) [current: ]"
• If yes: show the configurable sub-fields before asking each one:

  Watchdog sub-fields (press Enter to keep current value):
    stale_factor           — missed-cycle tolerance multiplier (e.g. 2)                   [current]
    escalate_after         — consecutive stale cycles before escalation (e.g. 3)          [current]
    operator_grace         — silence window before alert fires (e.g. 15m, 1h)             [current]
    context_clear_tokens   — emergency /clear when prompt tokens exceed this (e.g. 700000, 0=off) [current]
  

  Then ask each field in sequence.
• Write each changed field through settings-edit ... set watchdog.<field> <value> (watchdog.enabled, watchdog.stale_factor, watchdog.escalate_after, watchdog.operator_grace, watchdog.context_clear_tokens). Per-field dotted sets preserve any untouched siblings.
  • Note: "Changes take effect on the next watchdog run. To register or remove the OS timer: bin/hermit-watchdog install / bin/hermit-watchdog uninstall. Docker hermits run the watchdog from the entrypoint loop — no install step needed."
• Context hygiene compact (context_hygiene.compact — runs independently of the "Enable watchdog?" answer above, same as context_clear_tokens): ask "Enable routine-hygiene compaction? (yes / no) [current: ]". If yes, show the sub-fields:

  Context hygiene compact sub-fields (press Enter to keep current value):
    min_context_tokens     — routine-hygiene /compact when prompt tokens exceed this (e.g. 150000) [current]
    min_interval           — minimum time between compacts, avoids summary-of-summary loss (e.g. 4h) [current]
  

  Then ask each field in sequence. Write each changed field through settings-edit ... set context_hygiene.compact.<field> <value> (context_hygiene.compact.enabled, context_hygiene.compact.min_context_tokens, context_hygiene.compact.min_interval). No restart/reconcile step needed — the watchdog reads config.json fresh on every scheduler tick.

If argument is "routines":

• Show current routines from config.routines array:

  Routines (config.json routines → routine monitor; CronCreate fallback where Monitor is unavailable):
  
    #  ID           Schedule       Skill                                Status
    1. morning      30 8 * * *     claude-code-hermit:brief --morning    enabled
    2. evening      30 22 * * *    claude-code-hermit:brief --evening    enabled
    3. weekly-deps  0 9 * * 1      claude-code-hermit:session-start ...  disabled
  
  (or "No routines configured" if empty)
  

• Ask: "Add / edit / remove / enable / disable? (or 'done')"
• Add wizard: ask for:
  • ID (unique name, e.g., "weekly-deps")
  • Schedule — offer common presets, or accept raw 5-field cron:
    • Daily at 08:30 → 30 8 * * *
    • Weekdays at 09:00 → 0 9 * * 1-5
    • Sundays at 23:00 → 0 23 * * 0
    • Every 15 minutes → */15 * * * *
    • 1st and 15th of month at 10:00 → 0 10 1,15 * *
    • Custom → operator types raw cron
  • Skill to run (full slash-command name, e.g. claude-code-hermit:brief for plugin skills, ha-refresh-context for local project skills)
  • Enabled (yes/no, default yes)
  • Write to config.json routines array.
• Edit: select by number, change any field.
• Remove: select by number, delete from array.
• Enable/disable: select by number, toggle enabled field.
• Loop until operator says "done".
• After all edits are written, invoke /claude-code-hermit:hermit-routines load via the Skill tool to apply the new schedule live (no restart). Surface the result inline:
  • Success: "Routines reloaded: , ( total). Active immediately."
  • Failure: "Settings saved to config.json, but /claude-code-hermit:hermit-routines load failed: . Run /claude-code-hermit:hermit-routines load manually to apply."

If argument is "idle":

• Show current idle_behavior value
• Ask: "What should the hermit do when idle between tasks?
  1. Discover — run reflection and priority alignment (default)
  2. Wait — only check for new tasks and channel messages Choose 1-2: [current value]"
• Run settings-edit ... set idle_behavior <wait|discover>.

If argument is "env":

• Show current env values from config.json in a table:

  Environment Variables (config.json env → .claude/settings.local.json)
  
    AGENT_HOOK_PROFILE              standard
    CLAUDE_AUTOCOMPACT_PCT_OVERRIDE 65
    MAX_THINKING_TOKENS             10000
  

• Protected keys that cannot be changed via this command: AGENT_HOOK_PROFILE. These are managed by the boot script and docker-setup. If the operator tries to set one, respond: "AGENT_HOOK_PROFILE is managed by the boot script (standard for interactive, strict for Docker). To change it, edit config.json directly — the boot script validates on next start."
• Ask: "Set, change, or remove an env var? (e.g., 'MAX_THINKING_TOKENS 20000', 'remove MAX_THINKING_TOKENS', or 'done') [done]"
• Loop until operator says "done", "skip", or presses Enter:
  • If input targets a protected key: reject with the message above
  • If input is remove <KEY>: delete the key from env
  • If input is <KEY> <VALUE>: set env[KEY] = VALUE
• Note: "Env changes are written to .claude/settings.local.json on next hermit-start. To apply now, restart the hermit session."

If argument is "compact":

• Show current compact values from config.json:

  SHELL.md Compaction (config.json compact → session-archive.ts idle transition)
  
    monitoring_threshold    30    (compact when Monitoring exceeds this many lines)
    monitoring_keep         20    (keep this many recent entries after compacting)
    summary_threshold       30    (compact when Session Summary exceeds this many lines)
    summary_keep            15    (keep this many recent entries after compacting)
  

• Ask: "Change a threshold? (e.g., 'monitoring_threshold 50', 'summary_keep 20', or 'done') [done]"
• Loop until operator says "done", "skip", or presses Enter:
  • Validate: value must be a positive integer
  • Validate: *_keep must not exceed its corresponding *_threshold (setting keep equal to threshold effectively disables compaction for that section)
  • Update compact[key] in config.json
• Note: "Compaction runs at each idle transition (task completion). No restart needed."

If argument is "docker":

• Show current docker.packages list:

  Docker Packages (config.json docker.packages → Dockerfile.hermit)
  
    build-essential
    ffmpeg
  
  (or "No packages configured" if empty)
  

• Ask: "Add or remove packages? (e.g., 'add ffmpeg imagemagick', 'remove ffmpeg', or 'done') [done]"

• Loop until operator says "done", "skip", or presses Enter:

  • If input is remove <PKG> [<PKG>...]: remove the packages from docker.packages
  • If input is add <PKG> [<PKG>...]: add the packages to docker.packages (deduplicate)
  • If input is just package names without add/remove prefix: treat as add

• After changes, note: "Rebuild your container to apply: docker compose -f docker-compose.hermit.yml build"

• Then show current docker.recommended_plugins:

  Recommended Plugins (config.json docker.recommended_plugins)
  
    [enabled]  claude-code-setup (claude-plugins-official) — auto-installed on boot
    [enabled]  claude-code-homeassistant-hermit (claude-code-homeassistant-hermit) — auto-installed on boot
  
  (or "No recommended plugins configured" if empty)
  

  Display each entry as [enabled/disabled] <plugin> (<marketplace>) — show the org/repo (the marketplace field) in parens.

• Ask: "Enable, disable, add, or remove recommended plugins? (e.g., 'enable claude-code-setup', 'add claude-code-setup', 'add superpowers obra/superpowers-marketplace', 'remove superpowers', or 'done') [done]"

• Loop until operator says "done", "skip", or presses Enter:

  • enable <PLUGIN>: set enabled: true on matching entry
  • disable <PLUGIN>: set enabled: false on matching entry
  • remove <PLUGIN>: remove the entry entirely
  • add <PLUGIN> [<MARKETPLACE>]: add new entry with scope: "project", enabled: true. <MARKETPLACE> is an org/repo (e.g. obra/superpowers-marketplace) or omitted (defaults to anthropics/claude-plugins-official). If <MARKETPLACE> is provided but not registered locally, prompt: "Marketplace <MARKETPLACE> is not registered locally. Add it with claude plugin marketplace add <MARKETPLACE> first, then re-try." Abort the add. Dedupe rule: refuse the add if an existing entry has the same (plugin, marketplace) pair (scope is NOT part of the key) — operator should enable or remove first.
  • If input is just a plugin name without a verb: treat as enable if it exists, add if it doesn't

• After changes, note: "Restart container to install new plugins: .claude-code-hermit/bin/hermit-docker restart"

If argument is "scheduled-checks":

• Read state/reflection-state.json for runtime state (last run dates). If missing, show "(no runs yet)" for all.
• Show current scheduled_checks entries from config.json:

  Scheduled Checks (config.json scheduled_checks)
  
    #  ID                      Plugin               Trigger   Interval  Last Run    Status
    1. automation-recommender  claude-code-setup     interval  7 days    2026-04-01  enabled
    2. md-audit                claude-md-management  interval  7 days    (never)     enabled
    3. md-revise               claude-md-management  session   —         2026-04-06  enabled
  
  (or "No scheduled checks configured" if empty)
  

• Ask: "Enable, disable, add, remove, or change interval? (e.g., 'disable md-audit', 'interval automation-recommender 14', 'add my-check my-plugin /my-plugin:my-skill interval 7', 'add my-check my-plugin /my-plugin:my-skill session', or 'done') [done]"
• Loop until operator says "done", "skip", or presses Enter:
  • enable <id>: set enabled: true on matching entry
  • disable <id>: set enabled: false on matching entry
  • interval <id> <days>: update interval_days on matching entry (only valid for trigger: "interval")
  • add <id> <plugin> <skill> interval [days]: add interval-triggered entry with enabled: true, interval_days (default: 7). Deduplicate by id.
  • add <id> <plugin> <skill> session: add session-triggered entry with enabled: true. Deduplicate by id.
  • remove <id>: delete the entry from config and its state from state/reflection-state.json
• Note: "Interval checks run during idle reflection. Session checks run at task completion. Changes take effect on the next cycle."

If argument is "quality-gate":

Interactive (terminal) turn: Ask the operator via AskUserQuestion to pick a tier. Show the current value in brackets if quality_gate.tier is set.

Prompt: "Quality-gate tier for accepted-proposal auto-implementations. Controls whether /claude-code-hermit:simplify (cleanup pass) runs at step (e.5) of /proposal-act."

Options:

• Budget (default; recommended): /claude-code-hermit:simplify never runs. Cheapest. No post-implementation cleanup.
• Balanced: make an inline RUN/SKIP decision on each implementation from the proposal category and touched files (no subagent) — RUN triggers /claude-code-hermit:simplify, SKIP doesn't. Costs an occasional ~$0.25 /claude-code-hermit:simplify run when the decision is RUN.
• Quality: /claude-code-hermit:simplify runs on every implementation, no judgment. ~$0.25-$0.35 per implementation in Sonnet pricing.

Run settings-edit ... set quality_gate.tier <chosen> (creates the quality_gate object if missing; a legacy enabled sibling is preserved untouched — skill behavior reads tier only).

Channel-tagged turn: send the same prompt via the channel reply tool with the three tiers numbered (Budget/Balanced/Quality, same descriptions as above), AND queue a pending micro-proposal entry per reflect § Queuing procedure: options: ["budget", "balanced", "quality"], tier: 1, on_resolve: "/claude-code-hermit:hermit-settings quality-gate --answer {answer}". If invoked as quality-gate --answer <tier> (channel-responder resolving that entry), skip the ask and run settings-edit ... set quality_gate.tier <tier> directly, then confirm via channel.

Note: if you have claude-code-dev-hermit:dev-quality installed and you commit autonomous-implementation diffs through it, consider Budget — /dev-quality already runs /claude-code-hermit:simplify before commit, and any non-Budget tier here would double-fire the cleanup pass (~$0.40-$0.70 of duplicated spend per committed implementation).

If argument is "reflection":

• Show current value from config.json:

  Reflection (config.json reflection)
  
    graduation_min_sessions   1   (minimum distinct sessions before a ledger pattern graduates to a proposal candidate)
  

  (substitute the actual value; show 1 if absent)
• Ask: "Minimum distinct sessions before a pattern graduates? 1 = surfaces after first session (default, ships enabled); 2 = requires recurrence across two sessions (pre-v1.2.6 behavior). [current: ]"
• Accept a positive integer (≥1) as input. Run settings-edit ... set reflection.graduation_min_sessions <int> (creates the reflection object if missing).
• Note: "Takes effect on the next reflect run. Set to 1 for fast feedback on fresh hermits; dial to 2 if the operator channel gets noisy."

If argument is "push-notifications": Ask: "Send a PushNotification (desktop notification in your terminal app, plus mobile push if Remote Control is connected) on proactive alerts? Fires when no channel is enabled OR a configured channel is unreachable (missing pairing, empty allowed_users, all-disabled). In always-on Docker or headless tmux only the Remote Control mobile push will be visible. Note: push is one-way; operator-→hermit replies (micro-proposals, session recovery) require a channel. on — enable push notifications off — disable push notifications [current: ]" Run settings-edit ... set push_notifications <true|false> ("on" → true, "off" → false; or toggle push_notifications to flip).

If argument is "artifact-dashboard": Ask: "Publish a Hermit Dashboard artifact (status, proposal queue, weekly evolution) to a private claude.ai URL? Refreshed by /brief, /weekly-review, /proposal-create, and /proposal-act — no extra cost beyond those turns. Requires a /login-authenticated session with Artifacts entitled (Pro/Max/Team/Enterprise); on any other surface or without entitlement it silently falls back to the existing markdown-only delivery. Private to your account unless your org has Team/Enterprise Share. on — enable the dashboard artifact (default) off — disable [current: ]" Run settings-edit ... set artifacts.dashboard <true|false> (creates the artifacts object if missing; "on" → true, "off" → false; or toggle artifacts.dashboard to flip).

If argument is "artifact-proposals": Ask: "Publish a Proposals-page artifact (full text of open proposals, deep-linked from new-proposal announcements) to a private claude.ai URL? Refreshed by /proposal-create and /proposal-act — no extra cost beyond those turns. Same entitlement/fallback/privacy notes as the dashboard artifact above. on — enable the proposals-page artifact (default) off — disable [current: ]" Run settings-edit ... set artifacts.proposals <true|false> (creates the artifacts object if missing; "on" → true, "off" → false; or toggle artifacts.proposals to flip).

If argument is "artifact-weekly-review": Ask: "Publish a Weekly-review artifact (the compiled weekly report as a stable-URL page) to a private claude.ai URL? Refreshed by /weekly-review — no extra cost beyond that turn. Same entitlement/fallback/privacy notes as the dashboard artifact above. on — enable the weekly-review artifact (default) off — disable [current: ]" Run settings-edit ... set artifacts.weekly_review <true|false> (creates the artifacts object if missing; "on" → true, "off" → false; or toggle artifacts.weekly_review to flip).

If argument is "artifact-authorization": This records a decision only — it never runs apply-settings.ts and never touches a settings file from this session. A channel reply may only flip hermit config, never permissions (auto-mode classifier invariant); the actual grant is applied by hermit-start's boot-time applyArtifactGrant, outside any session. Ask: "This hermit publishes status/proposal/weekly-review pages via Claude Code's Artifact tool. Unattended sessions can't answer a permission prompt, so authorize publishes now, or bank the first publish of each enabled page yourself instead?

1. Authorize — grant applied automatically at next boot
2. Bank first publishes — you publish the first version of each page now; refreshes then reuse the same URL [current: <artifacts.publish_authorized value>]" On answer "Authorize" (or "on"/"yes"): run settings-edit ... set artifacts.publish_authorized true. Reply: "Recorded: artifact publish authorized. The grant (permissions.allow Artifact + auto-mode seed) is applied automatically at next boot — .claude-code-hermit/bin/hermit-stop then hermit-start to apply now. No settings files were modified from this session." On answer "Bank first publishes" (or "off"/"no"/"decline"): run settings-edit ... set artifacts.publish_authorized false. Reply: "Recorded: no standing grant. First publish of each enabled page must happen in an attended session (docs/artifacts.md § refresh procedure); refreshes then reuse the same URL without prompting." Channel re-entry: if invoked as artifact-authorization --answer "<label>" (channel-responder resolving a micro-proposal queued by hermit-evolve's Step 10 deferred-migration relay, per the CHANGELOG's artifact-publish-authorization instruction), skip the Ask above and match <label> case-insensitively by prefix against Authorize / Bank first publishes, then run the matching settings-edit command and reply exactly as above.

3. Write config

Scalar/enum branches already persisted their change via settings-edit (see step 2). For the branches that manipulate arrays or delete keys (channels, routines, env, compact, docker, scheduled-checks, brief) — which settings-edit can't express — write the updated config back to .claude-code-hermit/config.json directly. Confirm the change to the operator.