Multi-agent systems need explicit role contracts and handoff records so authority, context, open risks, acceptance, and recovery behavior travel with delegated work.
Who it is for: Orchestrator builders, agent-organization designers, and operators running two or more agents that share artifacts or decisions.
Begin with role charters to assign authority, then use handoff envelopes for transfers. Add workflow specifications for shared task graphs and recovery playbooks for failure and escalation behavior.
Begin with role charters to assign authority, then use handoff envelopes for transfers. Add workflow specifications for shared task graphs and recovery playbooks for failure and escalation behavior.
Choose a bundle when: several exact members independently fit. Start with Multi-Agent Operations Pack, or use Agent Tool Permissions vs Safety Gates for a guided comparison.
The product type and summary below show how each option differs. Open a product page for exact scope, tiers, buyer fit, and bundle membership.
A written role system for multi-agent setups: role-authority contracts, authority maps with one-owner-per-decision arithmetic, escalation matrices, and worked charters.
A written protocol for work passing between agents (or agents and humans): handoff envelopes covering context/authority/open-risks/acceptance/disposition, acceptance tests, and the acceptance-transfers-ownership rule.
A written method for turning fuzzy business workflows into precise, agent-executable specs: six compilation passes, a workflow-spec JSON schema, and machine-checkable validation checklists.
A written failure-recovery system for multi-step agents: 13-class failure taxonomy with detection signatures, decision tree, repair paths, and safe retry policy.
A written permission system for agent tool use: risk tiers, deny-by-default matrices, allowlist/blocklist policies, and an audit-log schema.
The contracts and envelopes are transport-neutral. They must be loaded into or exchanged through the orchestration and communication tools already used by the team.
No multi-agent runtime, message transport, scheduler, or automatic authority enforcement is included.
