Configuration

All configuration lives in the profile's cordis.patch.yml (or a home patch). The plugin exports no editable volatile fields, so Settings shows no generic form; Settings → Dreaming is a read-only observability page. Defaults are defined in lib/config.js. The most useful keys:

Key	Default	Meaning
`enabled`	`true`	Master switch
`dataDir`	`''`	Data directory; empty = `$DSH_HOME/dreaming`
`probe.enabled`	`true`	Predict on skill load
`probe.mode`	`'inline'`	`inline` (recommended, cheap) or `subagent`
`probe.sampleRate`	`1`	Sampling rate; known-gap skills are always probed
`probe.attachDeadlineMs`	`1500`	How long the skill result waits for the probe; on timeout only history is attached
`probe.maxPerHour`	`120`	Hourly probe cap
`judge.enabled`	`true`	Judge after the turn
`observe.maxTurnExtensionMs`	`600000`	Hard cap on the observation window
`prompt.awarenessSection`	`true`	Inject the global awareness section
`prompt.annotateSkillResult`	`true`	Annotate the skill result
`prompt.refreshMinutes`	`10`	Minimum refresh interval for the global section (KV-cache friendly)
`dream.enabled`	`true`	Nightly dream
`dream.time`	`'03:30'`	Local wall-clock time HH:MM
`dream.timeZone`	`''`	IANA zone; empty uses the process zone
`dream.scoreThreshold`	`78`	Below this score a judged episode is a mismatch case
`dream.autoApply`	`false`	`false` records suggestions only; `true` applies validated edits with backup and revert
`dream.editableSources`	`['user-dsh','user-agents']`	Skill sources that may be edited on disk
`dream.dailyTokenBudget`	`60000`	Daily dream token budget
`run.maxConcurrent`	`2`	Concurrent probe/judge calls
`store.retainDays`	`120`	Episode retention
`store.maxEpisodes`	`4000`	Episode cap

Example (probe office skills only and opt into automatic edits):

- id: dreaming
  config:
    probe:
      sampleRate: 0.5
      attachDeadlineMs: 1200
    dream:
      time: '04:10'
      timeZone: Asia/Shanghai
      autoApply: true

dsh-plugin-dreaming

安裝

說明文件

Configuration