Team Collaboration with Paperclip: Shared Agents and Permissions

AI agents aren’t built by one person. They’re built, maintained, and monitored by teams. Here’s how to manage Paperclip agents across your organization.

The Collaboration Problem

Without proper collaboration tools, teams face:

• Agents owned by individuals who leave
• No visibility into what agents are doing
• Uncoordinated changes that break production
• No audit trail for compliance

HostAgentes solves these with built-in team features.

Team Dashboard

Every team gets a shared dashboard showing:

• All agents — every agent across the organization
• Status overview — which agents are healthy, degraded, or down
• Usage metrics — request volume, token usage, costs by agent
• Recent activity — deployments, configuration changes, alerts

No more asking “who deployed this agent?” or “why is the bill so high?”

Role-Based Access Control

Roles

Role	Can View	Can Edit	Can Deploy	Can Manage Team
Viewer	Yes	No	No	No
Editor	Yes	Yes	Yes	No
Admin	Yes	Yes	Yes	Yes

How Teams Use Roles

Engineering team:

• Engineers are Editors (configure and deploy agents)
• Tech lead is Admin (manages team and billing)

Support team:

• Agents are owned by support managers (Editors)
• Agents monitor as Viewers
• Escalation rules configured by Editors

Finance team:

• Finance lead is Viewer on all agents (cost monitoring)
• No edit or deploy access — just visibility

Agent Ownership

Every agent has an owner — the person responsible for it:

• Owner receives alerts and notifications
• Owner can transfer ownership when leaving the team
• Ownership is logged for audit purposes
• Multiple editors can work on the same agent

Ownership Transfer

When someone leaves the team:

1. Admin opens team settings
2. Selects departing member
3. Clicks “Transfer agents”
4. Chooses new owner
5. All agents transferred with full history preserved

Audit Logging

Every action is logged:

• Configuration changes (what changed, who changed it, when)
• Deployments and rollbacks
• API key creation and revocation
• Team membership changes
• Access to conversation logs

Audit logs are retained for 1 year (longer on Scale plan) and exportable for compliance.

Multi-Team Management

Larger organizations have multiple teams:

Team Isolation

Each team has isolated:

• Agent configurations
• Environment variables
• Conversation logs
• Usage metrics

Teams can’t see each other’s agents unless explicitly shared.

Shared Agents

Some agents need cross-team access:

• Support agent — shared with support and engineering
• Data agent — shared with analytics and product
• Content agent — shared with marketing and product

Share agents with specific teams and control their access level (view or edit).

Organization-Level Controls

Organization admins can:

• Set global policies (allowed models, max token limits)
• View aggregated usage and costs across teams
• Enforce security requirements (SSO, IP whitelisting)
• Manage billing centrally

Collaboration Workflows

Development Workflow

1. Developer creates agent in staging environment
2. Reviewer tests agent quality and approves
3. Admin promotes to production with one click
4. Team monitors production metrics

Incident Workflow

1. Alert fires for degraded agent
2. On-call receives notification
3. On-call reviews conversation logs
4. On-call rolls back configuration if needed
5. Post-incident review with audit trail

Cost Management Workflow

1. Finance reviews monthly cost dashboard
2. Identifies high-cost agents
3. Engineering optimizes (model routing, caching)
4. Finance tracks improvement in following month

Getting Started

Team features are available on the Pro plan. Organization-level controls and SSO are on the Scale plan.

See team plans →