30+ agents.
Installed, not trained.
Browse roughly thirty agents across eleven categories. Every one is a readable spec, a rubric, and an audit trail — you install it, review it, and own it, not a black box that learns off your data. Fork the ones that don’t quite fit. Publish your own as private, tenant-scoped catalogue entries.
Roughly thirty agents,
fourteen featured here.
Email Drafter
Writes outbound email in your voice using thread history, client tone, team style, and project state.
Conversational Workflow Builder
Describe what you want in plain English. The builder parses intent, maps your installed agents’ schemas, drafts the workflow graph, previews it, and ships it when you confirm.
Meeting Prep Brief
Packages a pre-meeting brief from CRM, calendar invites, thread history, and the last time your team spoke to these people.
Proposal Writer
Assembles proposals and statements of work from the brief, your commercial template, and the client’s prior engagements.
Account Researcher
Pre-call brief from CRM, news, filings, and your firm’s prior engagement notes.
Project Status
Weekly status from Jira, commits, and Slack — written in your house format.
Quality Reviewer
Scores an output against your agent’s rubric and flags anything a human reviewer should take a second pass at.
Onboarding Brief
Packages the kickoff pack for a new hire or new client from the five layers your team maintains.
Answer Desk
Answers internal questions from your wiki, docs, and threads — with citations, so the answer is checkable.
Client Report Builder
Turns a month of delivery artefacts — tickets, commits, sent work — into the client report your PM writes.
Release Notes Writer
Turns a release branch’s commits + issues into the notes your product team ships externally.
Retrospective Facilitator
Reads the sprint — tickets, Slack, the retro doc — and drafts the themes and actions before the team walks in.
Multi-Language Translator
Translates drafts into the languages your team actually ships in — in your voice, not a generic translation.
Policy & Procedure Drafter
Drafts internal policies and procedures grounded in the company layer — legal posture, compliance matrix, escalation rules.
Every agent declares what context layers it needs. Missing a layer? The agent surfaces it, never guesses.Read the thesis →
See it write an email you’d recognize as your own.
[2] project / artefacts / film-cuts
[3] project / open-questions / Q-11
voice match gate · escalations: none · awaiting approval
Describe the workflow.
Ship the spec.
The Conversational Workflow Builder is an installed agent. Tell it what you want — in plain English — and it parses your intent, reads your installed agents’ input and output schemas, drafts the workflow definition, previews it for confirmation, and saves it. The spec is the artefact; plain English is the authoring surface.
# You said:
# "When I upload meeting notes, run the Client Report Builder,
# then if the report mentions action items, draft an email for each one."
# Workflow Builder drafted:
name: Notes → Report → Follow-ups
trigger: upload # meeting-notes artefact
steps:
- run: client-report-builder
with: { notes: $trigger.artefact }
out: report
- if: report.action_items.length > 0
for_each: report.action_items
run: email-drafter
with: { topic: $item }
# Preview? [y/N] — saves to your workspace on confirm.Every agent product starts the same way — a black box that “learns your team” off data you never agreed to hand over. Ours are installed, not trained. A readable spec, a rubric, an audit trail. You can read them before you run them, and fork them when they don’t fit. That’s the bargain I want on the other side of the table.
