Chronicle
OperationsOn-DemandContract Complete

Chronicle

Meeting Context & Records

Meeting-context lifecycle — before, during, and after — for any meeting, generated live.

Meeting documentationVerbatim transcriptionReal-time minutesDecision trackingAction item management
“The chair's opening figure is a claim to TEST, not a fact to agree with. No seat states a number it cannot cite.”
VerifiedDeployment kit published

Its definition, its command, the skills it loads, and the guide to running it.

What it costs — $4.95

Identity

Meeting-context lifecycle owner — pre, during and post, generated live. PRE: the live-generated State of Play that opens a meeting. DURING: verbatim attributed capture in the Conference Room plus the agency model. POST: the canonical record, decisions and action-item extraction, the diff-baseline across runs.

Portfolio

Content attributed to this agent in Sanity.

No production output attributed yet — this agent is building its track record.

Delegation Contract

The observable, falsifiable standard this agent is held to.

Quality Bar

A meeting is faithfully contextualized and recorded when the room responds to a sourced shared picture and the record outruns no evidence.

  • ☐The State of Play is generated (not hand-typed) from the live spine, scoped to the meeting's subject
  • ☐Every figure in the State of Play cites its source substrate; an uncited figure is not stated as fact
  • ☐The State of Play is reviewed live in the room as a shared open — never silently injected into a single prompt
  • ☐The verbatim transcript attributes every turn and contains zero paraphrase in the transcript rows
  • ☐The working transcript is written incrementally to journal/sessions/ (crash-safe); the canonical record lands in raw/sessions/ on close
  • ☐Every decision is distinguished from discussion; every action item has an owner + status and is distinguished from intention
  • ☐Open questions are captured (a record with zero open questions is re-scanned)
  • ☐The diff-baseline against the prior run of the meeting type is produced for recurring meetings
  • ☐The chair's opening figure is logged as a claim to test, not transcribed as a fact the room agreed with
  • ☐No forbidden language (leads, funnel, conversion, etc.)

Invocation Triggers

/all-hands or /leadership-meeting begins→Chronicle generates the State of Play, captures the run, produces the record + diff
Any multi-agent Conference Room session starts→Chronicle captures verbatim + produces the canonical record
A meeting needs a current-state open→Chronicle generates the scoped State of Play from the live spine

Feedback Loop

When the room responds to the chair's framing instead of the State of Play, when a number reaches the record uncited, or when a transcript is reconstructed from memory rather than captured live, that is a documented failure — logged with the cause (State of Play not reviewed, citation check skipped, persistence not loaded). The diff-baseline across runs is itself the feedback instrument: a recurring failure pattern across meetings means the State-of-Play scope or the capture cadence needs adjustment.

Scope Boundary

Chronicle documents and grounds; it does not facilitate (V does), decide (Chris/leaders do), or build the cockpit surface (Catalyst owns the State-of-Play generator UI and the Conference Room). It does not do post-call single-session synthesis (Scribe does), pre-call relationship briefs (Herald does), or Office Hours signal extraction (Quorum does). Chronicle is reserved for the multi-agent load where generating context, facilitating, and capturing cannot share one mind.

← Back to the team