{"name":"Operator Commons","description":"The trust, consent and discovery layer above MCP and A2A: operators publish playbooks and kits describing how their AI agents work, then share them with operators they trust under explicit, revocable consent grants. Indexes and exchanges context; never installs or executes anything on an operator's machine.","supportedInterfaces":[{"url":"https://operatorcommons.ai/api/mcp/messages","protocolBinding":"MCP","protocolVersion":"1.0"},{"url":"https://operatorcommons.ai/api/a2a/v1","protocolBinding":"HTTP+JSON","protocolVersion":"1.0"},{"url":"https://operatorcommons.ai/api/a2a/playbook-export","protocolBinding":"HTTP+JSON;profile=operator-commons-playbook-export","protocolVersion":"1.0"}],"provider":{"organization":"Keen Media Inc.","url":"https://operatorcommons.ai"},"version":"2.24.3","documentationUrl":"https://operatorcommons.ai/llms.txt","iconUrl":"https://operatorcommons.ai/brand/app-icon-512.png","capabilities":{"streaming":false,"pushNotifications":false,"extendedAgentCard":false},"securitySchemes":{"bearer":{"httpAuthSecurityScheme":{"scheme":"Bearer","bearerFormat":"act_<opaque>","description":"Operator-issued agent token from /settings/tokens. Stored SHA-256-hashed, scope-checked on every request."}},"shareId":{"apiKeySecurityScheme":{"location":"query","name":"shareId","description":"Export-only capability link issued from the share UI, accepted by the playbook-export interface and no other. Honors the playbook's share grant and is audit-logged."}}},"securityRequirements":[{"schemes":{"bearer":{"list":[]}}}],"defaultInputModes":["application/json","text/plain"],"defaultOutputModes":["application/json","text/markdown"],"skills":[{"id":"playbook_export","name":"Playbook export","description":"Export an operator's playbook as structured JSON over HTTP, authorized by a scoped agent bearer token or by a shareId capability link that honors the playbook's share grant.","tags":["playbook","export","consent"],"examples":["Export the playbook I was granted access to as JSON."],"securityRequirements":[{"schemes":{"bearer":{"list":["playbook:export"]}}},{"schemes":{"bearer":{"list":["playbook:export:any"]}}},{"schemes":{"shareId":{"list":[]}}}]},{"id":"setup_comparison","name":"Agent setup comparison","description":"Compare two agent setups or playbooks and report the differences — rules, skills, subagents, commands, MCP servers, hooks and routines — so an agent can see what a trusted operator does differently.","tags":["comparison","playbook","kit"],"examples":["Compare my agent setup with the kit I was granted access to."],"securityRequirements":[{"schemes":{"bearer":{"list":["playbook:compare"]}}},{"schemes":{"bearer":{"list":["playbook:export"]}}},{"schemes":{"bearer":{"list":["playbook:read"]}}},{"schemes":{"bearer":{"list":["setup:compare"]}}}]},{"id":"setup_share_negotiation","name":"Consent-gated setup sharing","description":"Negotiate access to another operator's setup: request scopes, let the owner's sharing policy auto-approve or route to human review, and read sanitized projections through the resulting scoped, expirable, revocable grant.","tags":["consent","grants","trust"],"examples":["Request read access to @alex's sanitized setup summary."],"securityRequirements":[{"schemes":{"bearer":{"list":["sanitized_setup_summary"]}}},{"schemes":{"bearer":{"list":["setup_share:grants:list"]}}},{"schemes":{"bearer":{"list":["setup_share:request"]}}},{"schemes":{"bearer":{"list":["setup_share:status:read"]}}},{"schemes":{"bearer":{"list":["setup_snapshot:read"]}}}]},{"id":"context_sanitization","name":"Pre-publish content scan","description":"Advisory scan of agent context before it is published or shared: detects secret-shaped values (API keys, access tokens, private keys), sensitive vocabulary (credentials, personal, financial, medical, legal, HR and client terms) and soft signals such as email addresses and private IP addresses, and returns a verdict with a suggested risk level. It never rewrites the input or returns a cleaned copy; the caller must remove what the verdict flags. Enforcement is separate and server-side: secret-shaped values are rejected at publish, and never-share categories (secrets, environment values, tokens, private memory, local absolute paths, client context) cannot be granted to another operator whatever scopes are requested.","tags":["privacy","sanitization","safety"],"examples":["Scan this configuration snippet for secrets before I publish it."],"securityRequirements":[{"schemes":{"bearer":{"list":[]}}},{"schemes":{}}]},{"id":"workflow_recipe_publish","name":"Workflow recipe publishing","description":"Author and publish structured playbook sections — workflow recipes, review loops and verification rules — validated against the published playbook schema before they are stored.","tags":["playbook","authoring","publishing"],"examples":["Publish this review-loop recipe to my playbook."],"securityRequirements":[{"schemes":{"bearer":{"list":["playbook:publish"]}}},{"schemes":{"bearer":{"list":["playbook:write"]}}}]},{"id":"a2a_message_send","name":"Ask a trusted operator's agent","description":"Send an A2A message that asks a named operator's agent for setup-share access. Creates a consent request evaluated against the target's sharing policy; returns a Task whose terminal state carries the resulting grant or the denial reason. Name the operator with the interface tenant or targetHandle in an application/json data part, alongside scopesNeeded.","tags":["a2a","consent","hosted-delegate"],"examples":["Ask @alex's agent for read access to their sanitized setup summary."],"securityRequirements":[{"schemes":{"bearer":{"list":["trust:invite"]}}}]}]}