seeded meeting-prep fixture
fixture seed until real packetObserve -> Workflow IR -> Skill Pack -> Improve
Apprentice learns one real workflow and turns it into a reusable skill.
YC Proof Path
Session to reusable skill
1 candidates; api native recommended
useful after editsmeeting-prep-agent skill pack: 9 files, 5 eval cases
skill pack export ready0 saved rerun records
real replay pending19 proposals, 4 next eval cases
operator review requiredWorkflow IR, skill pack export, supervised reuse, and draft improvements run in app.
3 demo profiles, 0 verified partners.
Real partner-approved replay still needs a packet path and proof-gate run.
Current demo data is explicitly fixture-backed. The YC proof gate is a real partner-approved redacted packet replayed without fixture mode.
Needs Attention
Workflow operations queue
manual context stitching, meeting prep pattern, follow up drafting pattern, connector overlap, repeated app sequence; raw upload blocked
partner-meeting-prep-and-follow-up-candidateNo audit log exists for this registered workflow.
workflow_weekly_investor_updateNo audit log exists for this registered workflow.
workflow_candidate_sourcing_follow_upNo audit log exists for this registered workflow.
workflow_team_digestNo audit log exists for this registered workflow.
workflow_research_briefusefulness 71; confidence 91; risk 29; missing: operator-confirmed workflow goal and source-of-truth notes, explicit decision points from the operator
observer_partner_meeting_prep_2026_04_28Guided UI vs compiled workflow
Competitive demo
A UI guide helps the user complete one task; Apprentice turns the same session into a reusable, governed agent skill.
An AI-native operator has a recurring partner meeting. Today they manually open Calendar, Gmail, Slack, Drive, and notes to prepare a brief and draft follow-up.The operator clicks Record Workflow and receives a Clicky-like prompt while opening the calendar invite and searching related Gmail and Slack context.
The audience recognizes the familiar guided-through-UI pattern in the first 20 seconds.
Apprentice shows the inferred Workflow IR from its internal Workflow Compiler engine: trigger, inputs, apps, data sources, steps, decision points, sensitive data, and scores.
The audience can explain what work should no longer require manual UI traversal.
The dashboard contrasts browser/computer-use fallback with API, MCP, and skill-first execution, then selects the meeting-prep agent/tool path.
The audience sees that Apprentice is compiler-backed, not a desktop copilot clone.
The same captured workflow becomes a meeting-prep artifact bundle, follow-up draft policy, MCP tool stub, eval checklist, and registry promotion candidate.
A design partner asks to reuse the workflow for the next recurring meeting.
The approval panel shows read-only, draft-only, and approval-required modes. Follow-up messages are drafts; external sends stay blocked.
The audience trusts the workflow enough to approve a draft-only recurring prep run.
The final screen shows the workflow queued for registry review so the next matching meeting can produce a brief without repeating the UI path.
The design partner says they would stop doing this workflow manually.
Partner Replay Intake
Review redacted replay packet
Local parse and sanitized metadata only
No raw packet JSON is stored in localStorage or sent to backend payloads.Import only a partner-approved redacted packet. Fixture packets fail by default in this UI.
Usefulness verification
Observer Review
Local deterministic review loaded. Packet source: seeded meeting-prep fixture.
Backend persistence: No backend observer outcome has been saved yet. Saved records this session: 0.
71/100
local-workflow-observer-v091/100
Capture evidence covers enough apps/events to review the workflow boundary.29/100
Observer packet excludes raw screen video, raw typed text, private message bodies, and full browsing history.needs feedback
operator accepted, edited, reused, or observed time saveddecision points: Add explicit decision questions for attendee matching, evidence sufficiency, and follow-up readiness.
- Add an explicit source-of-truth order for Calendar, Gmail, Slack, Drive, and manual notes.
- Add a stop condition when attendee/company matching is ambiguous.
- Expose a dry-run mode that returns planned connector queries and missing context.
- Return evidence_refs and unsupported_claims as first-class output fields.
- Require approval for expanded connector scopes, new recipients, recurring schedule changes, and any external send.
- Keep raw capture discard visible before registry promotion.
- Generated artifact is accepted or edited by the operator before registry promotion.
- At least one reuse run saves measurable time without unsupported claims.
- Which source should win if Gmail and Slack disagree?
- What exact brief sections do you reuse before every meeting?
- What would make the follow-up draft unsafe to send?
Skill Pack Compiler
Skill Studio
v0.1.0
SKILL.md entrypoint
85 skill vs 52 baseline
No raw capture or private content included
Skill pack manifest for export, reuse, and registry review.
Codex-compatible skill entrypoint generated from Workflow IR.
Human-readable summary of the generated skill pack.
Durable Workflow IR object used to generate the skill.
MCP-style tool stub with schemas, permissions, and dry-run behavior.
API/tool-first retrieval plan for approved context sources.
Approval and blocked-operation rules for the workflow.
Generated and observer-suggested eval criteria.
Basic eval harness cases for the exported skill pack.
82 vs 58 baseline, delta 24
88 vs 52 baseline, delta 36
84 vs 47 baseline, delta 37
86 vs 50 baseline, delta 36
87 vs 55 baseline, delta 32
--- name: meeting-prep-agent description: Prepare a read-only meeting brief and draft follow-up from approved context. --- # Meeting Prep Agent ## Purpose Turn the Workflow IR "Partner meeting prep and follow-up" into a repeatable meeting-prep and follow-up drafting process for AI-native operators. ## When To Use Use when Calendar event with external attendee and company domain 30 minutes before start. ## Required Inputs - calendar invite - recent Gmail threads - Slack mentions - Google Drive docs - previous meeting notes - browser context from allowlisted pages ## Connector And Tool Preferences - Prefer Calendar, Gmail, Slack, and Drive APIs/connectors before browser or computer-use automation. - Use browser context only for allowlisted pages. - Use manual notes when the operator supplies context that connectors do not cover. ## Workflow Steps - 1. Inspect upcoming calendar event and attendee domains. - 2. Collect recent email and Slack context for attendee and company. - 3. Find relevant Drive docs and previous meeting notes. - 4. Generate brief, agenda, open questions, and commitments. - 5. Draft follow-up after meeting, requiring approval before external send. - 6. observe: Open the upcoming partner meeting invite and identify attendees. - 7. keyboard shortcut: Search email for the attendee domain. - 8. click: Open the most recent relevant email thread. - 9. wait for state: Wait for Drive meeting notes to load. - 10. type: Draft the safe meeting brief notes. - 11. Compile reviewed capture evidence into a Workflow IR draft before generating durable artifacts. ## Approval Rules - Default mode: draft_only. - External sends are blocked unless the user explicitly approves a separate send action. - Recurring runs, expanded scopes, new recipients, and state-changing writes require approval. - Do not request raw keystrokes or always-on screen recording. ## Output Format - One-page meeting brief - Attendee and company summary - Last contact and recent commitments - Open questions - Suggested agenda and asks - Evidence references - Follow-up draft only when the user provides notes or asks for a draft ## Eval Checklist - Trigger matches the intended calendar event and attendee/company context. - Required Calendar, Gmail, Slack, Drive, browser, and manual-note inputs are covered when available. - Output includes attendees, company context, last contact, open commitments, relevant docs, agenda, and suggested asks. - Evidence references support commitments, suggested asks, and follow-up claims. - Unsupported claims are absent. - Sensitive internal content is not overexposed. - Permission scopes match the approved manifest. - Approval policy is followed before any recurring run, expanded scope, write, or external send. - Rollback steps are available and understandable. - Artifact is reusable as a reviewed template after a dry run. - Follow-up recipients are correct. - Follow-up tone fits the relationship and meeting context. - No external follow-up is sent autonomously. ## Failure Handling - If missing relevant thread, stop and ask for review instead of guessing. - If wrong attendee match, stop and ask for review instead of guessing. - If overly broad context retrieval, stop and ask for review instead of guessing. - If capture evidence does not match the intended workflow boundary, stop and ask for review instead of guessing. - If UI state differs from the imported plan, stop and ask for review instead of guessing. - If selector target no longer exists, stop and ask for review instead of guessing. - If wait condition cannot be observed reliably, stop and ask for review instead of guessing. ## Non-Goals - Do not send external messages autonomously. - Do not operate as an always-on screen recorder. - Do not store raw keystrokes. - Do not expand beyond approved connector scopes.
No skill pack export has been saved yet.
Evidence-derived skill updates
Skill Improvement Loop
Operator review is required before any skill file changes.
meeting-prep-agent skill pack
passes basic eval at 100% eval pass rate.
No raw capture, private content, or autonomous mutation.
Observer, operator, or existing proposal evidence contradicts the current skill behavior, so the review drafts a skill revision for operator review.
Partner meeting prep and follow-up candidate: manual context stitching, meeting prep pattern, follow up drafting pattern, connector overlap, repeated app sequence.
useful after edits observer review with 3 skill suggestion(s).
Operator feedback 1: 1 reuse(s), 1 edited field(s), dismissed false.
Add an explicit source-of-truth order for Calendar, Gmail, Slack, Drive, and manual notes.
Add a stop condition when attendee/company matching is ambiguous.
Require evidence references next to commitments, suggested asks, and follow-up claims.
Expose a dry-run mode that returns planned connector queries and missing context.
Return evidence_refs and unsupported_claims as first-class output fields.
Keep external follow-up drafting as a separate approval-required operation.
Add an explicit source-of-truth order for Calendar, Gmail, Slack, Drive, and manual notes.
Require approval for expanded connector scopes, new recipients, recurring schedule changes, and any external send.
Brief includes attendees, company, last contact, open tasks, relevant docs, agenda, and suggested asks.
Add explicit decision questions for attendee matching, evidence sufficiency, and follow-up readiness.
Which source should win if Gmail and Slack disagree?
What exact brief sections do you reuse before every meeting?
Brief includes attendees, company, last contact, open tasks, relevant docs, agenda, and suggested asks.
Follow-up draft includes only facts supported by retrieved context.
No external message is sent without approval.
Every imported step maps to an ordered Workflow IR step
Curator keep pack: 22 min saved, eval 100%, edits -1
Curator revise pack: 4 min saved, eval 67%, edits 3
Curator archive pack: 0 min saved, eval 100%, edits 0
Curator consolidate pack: 22 min saved, eval 100%, edits -1
Curator pinned pack: 4 min saved, eval 67%, edits 3
Report skill_pack_curator_report_2026_04_28 is dry-run only; no skill packs were mutated, archived, consolidated, deleted, sent, or patched.
View curator reportThis is the implemented Apprentice improvement loop: observer/eval/reuse metadata becomes draft skill updates, eval cases, and knowledge entries. It is not a current AutoContext or GBrain dependency.
First proposal: SKILL.md - Updated skill pack passes the basic eval harness and still excludes raw capture/private content.
View API payloadView evidence reviewWorkflow Inbox
Captured operator workflows
Approve candidate, review Workflow IR diff, then promote only metadata and generated artifacts.
Workflow IR
Executive workflow
- Trigger
- Calendar event with external attendee and company domain 30 minutes before start
- Execution surface
- api native: Calendar, Gmail, Slack, Drive, Browser are available through approved connectors; outputs are read-only or draft-only and external sends are blocked.
- Inputs
- Calendar, Gmail, Slack, Drive, Browser
Connector Evidence
Normalized source events
Upcoming external partner meeting with company-domain attendee.
Recent email thread includes open question about integration timeline and owner.
Internal Slack thread mentions launch blocker and suggested agenda item.
Drive doc contains prior implementation notes and unresolved next steps.
Structured UI Step Plan
Imported plan normalized into Workflow IR
Imported partner meeting prep UI plan
5 ordered steps mappedbrowser automation
UI selectors are evidence, not execution authority.typed values redacted as metadata
Typed values stay redacted unless explicitly allowlisted later.2
imported-ui-plan-playwright-draftUse CUA Driver as a research reference for replayable trajectories and MCP-compatible computer-use fallback, but do not integrate it until API/MCP/CLI and Playwright surfaces fail a validated workflow.
Partner Evidence Ledger
Concierge sprint operating queue
5 partners / 2 paid signals
Targets are goals, not achieved traction.3
Visibly separated from partner-confirmed proof0
0 partner-confirmed signals0
0 until a verified payment record is linkedReplace fixture row with a partner-approved capture session before claiming traction.
demo fixture / 4/28/2026
Collect a real example recurring meeting and source metadata before upgrading this row.
demo fixture / 4/28/2026
Run intake interview and replace demo fixture with a partner-approved packet.
demo fixture / 4/28/2026
Verification
Permission, eval, and audit trail
calendar.read, gmail.read, slack.read, drive.read
Raw connector bodies are not rendered in the dashboard; raw content is summarized ephemerally and deleted after Workflow IR extraction unless explicitly retained.public or low sensitivity, email content, private message
No escalation flags in canonical fixture85/100
0 blocking failures2 metadata-only entries
brief_ref:partner-meeting-2026-04-30Workflow Registry
Reusable agent-native templates
Calendar event with external attendee and company domain 30 minutes before start
Friday digest window or explicit workflow capture
Candidate thread with unresolved next step
Daily digest window or explicit team-summary request
Manual research brief request with target company or topic
Team Workspace
Operator Apprentice Sprint
4
founder operator, chief of staff, admin, reviewer5
workflow_partner_meeting_prep, workflow_weekly_investor_update, workflow_candidate_sourcing_follow_upAudit export on
Keystrokes and always-on screen recording default to off.metadata only
ephemeral delete after irTeam Governance
Policies, exports, and analytics
3
read only, draft only, defer to checklistcalendar, gmail, slack, drive, browser
Raw capture TTL: 12 hoursjsonl metadata-only
audit-export://workspace_operator_team/2026-04/team_audit_export_2026_04.jsonl18 recurring runs
9 approved, 1 rejectedoperator-workflow-compiler
1 tool, 3 resources, 2 prompts/v1
5 metadata-only resources2 listings
Partner meeting prep and follow-up78% avg success
2 checklist degradationsGenerated Artifacts
Agent-ready outputs
Report Export
Agent Workflow Report bundle
Founder operator sprint
founder operator / Operator officePartner meeting prep and follow-up
api_native4.2 h/mo
18 min/run x 14 runsdraft only
Raw keystrokes, always-on screen, and autonomous sends are off.Calendar, Gmail, Slack, Drive, Browser are available through approved connectors; outputs are read-only or draft-only and external sends are blocked.
Approval Policy
Safe execution defaults
- Read approved connector context during prep windows.
- External messages remain blocked unless explicitly approved.
- 4 connector reads planned; raw private content is excluded from logs.
- Rollback preview: delete generated draft; disable scheduled run; revoke connector token.