Scribe
Operations On-Demand Fully Operational

Scribe

Transcript Processing Specialist

Transforms session transcripts into actionable intelligence

""A transcript is not a record of what happened. It is raw material for what happens next.""
01 · Scope

What is this agent's job?

Extracts intelligence from transcripts across eight dimensions and updates HubSpot engagement records.

Identity

Scribe transforms session recordings into compounding knowledge. Every transcript processed makes every future session better -- the synthesis feeds Herald's meeting prep, the signals feed Sentinel's drift detection, the content opportunities feed the publishing pipeline, and the contributor observations feed team intelligence. Scribe is the origin point for most of the org's intelligence.

Quality Bar

Every session synthesis extracts actionable intelligence across all 8 dimensions and is immediately usable by Herald, Sentinel, and the content pipeline.

  • ☐ Synthesis contains decisions, action items, traps, signals, breakthrough moments, relationship intelligence, delivery intelligence, and contributor observations
  • ☐ Every action item is trust-based (readiness criteria), not date-based
  • ☐ Relationship signals section populated with Value Path stage evidence
  • ☐ Content opportunities section present with methodology grounding
  • ☐ Synthesis saved to clients/{slug}/sessions/{YYYY-MM-DD}-{type}.md
  • ☐ HubSpot updates proposed with confirmed record IDs
  • ☐ No forbidden language

Invocation Triggers

"Process the transcript from..." or raw transcript file referenced → spawn scribe
New transcript appears in clients/{slug}/transcripts/ → spawn scribe
/daily-ops identifies unprocessed sessions → spawn scribe per session

Scope Boundary

Scribe processes transcripts into intelligence. Scribe does not schedule sessions, manage recordings, or publish articles.

What Works / What Doesn't

What Works

  • Comprehensive 8-dimension extraction process with 17-point quality checklist
  • Structured synthesis output with consistent format consumed by Herald, Sentinel, and content pipeline
  • HubSpot integration (appointment creation, engagement notes, task creation)
  • Special scoping mode for discovery sessions with structured requirements output

What Doesn't Work

  • Upstash indexing gap -- syntheses created locally but not automatically indexed for semantic search
  • Fully manual trigger -- no automation detects new transcripts and kicks off processing
  • No batch mode -- multiple sessions require individual triggers
Feedback Loop Post-session review: Herald's brief quality depends on Scribe's synthesis. Chris's corrections during /meeting-prep feed back as extraction quality signals. Echo captures repeated extraction misses in pattern-memory.json.
02 · Access

What can this agent touch?

Tools

ReadGrepGlobEditWriteBash

Model Tier

Opus Premium tier — synthesis and judgment

Gateway Routing

  • HubSpot writes via Ledger

Handoff

Herald (session prep), Sentinel (engagement signals), Forge (content multiplication)

03 · Context

What does this agent know?

Expertise Pack

Transcript Synthesis
skills/transcript-synthesis/

Startup Loads

  • agents/transcript-processor/AGENT.md
  • skills/enforcement/vf-platform-context.md
  • skills/enforcement/vf-self-correction.md
  • skills/transcript-synthesis/README.md
04 · Production Record

What has this agent produced?

Recent Runs

Run history coming soon — instrumentation in flight.

Active Engagements

HubSpot engagement attribution coming soon — created_by_agent stamping shipped today and will populate as new work is created.

Published Artifacts

No published artifacts attributed yet — this agent is building its track record.

Leadership Commentary

V (COO)
"Scribe is the origin point. Everything downstream -- Herald's briefs, Sentinel's enrichment, Forge's articles, Prism's profiles -- starts with a well-processed transcript. The quality of Scribe's extraction directly determines the quality of every agent that reads its output. The Upstash indexing gap is the most impactful infrastructure issue in the system: we have the query tools, the data, and the consumers, but the automatic connection between synthesis creation and searchability is missing."
← Back to AI Team