{"generated_at":"2026-10-14T20:07:22+00:00","item":{"added_at":"2026-09-19","analysis_method":"independent_manual_review_2026-10-10","assessment_status":"provisional","auth_posture":"auth_required","category":"Web Connector","confidence":"medium","description":"Send transactional email through verified domains with AgentiSend. Preflight messages, send individually or in batches, and inspect delivery outcomes. Manage sender budgets, approvals and account controls through API-key or OAuth access.","did":"0fbf0daf-cdd3-47f9-8093-20f15f0d5cbe","documentation_url":"https://agentisend.com/docs/guides/mcp","editorial_description":{"extended_description":"Batches contain up to 500 messages with per-item results. The source describes preflight as performing send checks without sending or charging. Additional tools cover contacts, suppressions, templates, received email, webhooks and domain verification; more operations are discoverable beyond the initial inventory.","schema_version":1,"summary":"Send transactional email through verified domains with AgentiSend.\nPreflight messages, send individually or in batches, and inspect delivery outcomes.\nManage sender budgets, approvals and account controls through API-key or OAuth access."},"homepage_url":"","icon":"https://raw.githubusercontent.com/fortuneflick/agentisend-claude-plugin/main/assets/logo.svg","id":"0fbf0daf-cdd3-47f9-8093-20f15f0d5cbe","installs":0,"last_scanned":"","license":"","long_description":"Batches contain up to 500 messages with per-item results. The source describes preflight as performing send checks without sending or charging. Additional tools cover contacts, suppressions, templates, received email, webhooks and domain verification; more operations are discoverable beyond the initial inventory.","manual_assessment":"high","manual_review":{"all_findings":[{"evidence":"The inventory explicitly exposes single, reply and batch email sends, with the listing allowing batches of up to 500 recipients.","reasoning":"These tools send external messages as the authenticated sender.","risk_type":"messaging_as_user","severity":"high","title":"Sends, replies to and batches transactional email","triggering_tools_or_skills":["reply_to_email","send_batch","send_email"]},{"evidence":"The inventory includes explicit deletion of contacts and webhooks.","reasoning":"These actions destroy persistent customer and integration resources.","risk_type":"data_deletion","severity":"high","title":"Deletes contacts and webhooks","triggering_tools_or_skills":["delete_contact","delete_webhook"]},{"evidence":"The connector can revoke an OAuth grant and set a sender limit; its description also identifies budget and kill-switch controls.","reasoning":"OAuth revocation and sender-control changes are security and administrative actions.","risk_type":"credential_admin","severity":"high","title":"Revokes OAuth grants and changes sender limits","triggering_tools_or_skills":["revoke_oauth_grant","set_limit","get_kill_switch_state"]},{"evidence":"The listing states that the hosted server exposes 81 tools, only 43 through tools/list and the rest through list_more_tools; the packet inventories only those 43.","reasoning":"The reachable action set is incomplete and can hide additional High operations.","risk_type":"generic_gateway","severity":"high","title":"Dynamically loads 38 undeclared tools","triggering_tools_or_skills":["list_more_tools"]},{"evidence":"The server exposes received emails and threads, which contain sender-controlled content.","reasoning":"Email is an untrusted prompt-injection surface.","risk_type":"prompt_injection_exposure","severity":"low","title":"Inbound email content is returned to the model","triggering_tools_or_skills":["get_received_email","list_received_emails","list_threads"]}],"assessment_status":"provisional","auth_and_scopes":"Bearer API key or OAuth 2.1 is described, with sender budgets and a kill switch; the scopes and the 38 actions dynamically exposed through list_more_tools are not enumerated.","blocked_missing_evidence":null,"confidence":"medium","displayed_risk":"high","evidence_completeness":"insufficient","manual_overall_risk":"high","official_sources":[{"accessed_at":"2026-10-10T20:24:37Z","url_or_path":"plutosecurity/plutonium@b749c8dc43536c794409be9805edd48063f045e2:app/data/conn/claudesec/0fbf0daf-cdd3-47f9-8093-20f15f0d5cbe.json","what_it_proves":"Candidate PR #125 record identity (name \"AgentiSend\", publisher, type web_connector) and 43 stored tool names; no tool descriptions stored"},{"accessed_at":"2026-09-22T10:46:07Z","url_or_path":"https://api.anthropic.com/api/directory/servers?verified_tier=anthropic,partner,community (snapshot claude-directory-all-tiers-2026-09-22.json)","what_it_proves":"Official Claude Directory listing https://claude.ai/directory/0fbf0daf-cdd3-47f9-8093-20f15f0d5cbe: complete tool_names inventory (43 tools), remote streamable-http endpoint https://api.agentisend.com/mcp, auth_posture auth_required, directory tier community"}],"primary_reason":"The declared tools already send email, delete resources and revoke grants, while 38 additional actions are hidden behind dynamic loading.","provisional_risk":null,"publication_recommendation":"fix","requires_runtime_verification":true,"review_key":"claude:0fbf0daf-cdd3-47f9-8093-20f15f0d5cbe","reviewed_at":"2026-10-10","reviewer_notes":"Reviewed all 43 Directory tool names. Despite candidate_tools_truncated=false, the listing itself says 81 tools exist; the candidate must enumerate and assess the 38 list_more_tools actions before publication.","risk_domains":{"arbitrary_network_access":"unknown","authz_ambiguity":"true","browser_control":"false","credential_or_admin_actions":"true","deletes_data":"true","deployment_or_infrastructure":"false","dynamic_hidden_actions":"true","filesystem_or_repo_access":"false","forwards_data_to_third_party":"true","modifies_data":"true","persistent_credentials":"true","prompt_injection_exposure":"true","provenance_uncertainty":"true","reads_private_data":"true","runs_code_or_queries":"unknown","sends_messages_as_user":"true","spends_or_moves_money":"unknown"},"tools_or_skills_reviewed":43,"tools_or_skills_total":43},"name":"AgentiSend","needs_runtime_verification":true,"plutonium_url":"https://plutonium.pluto.security/detail.html?planet=claudesec&did=0fbf0daf-cdd3-47f9-8093-20f15f0d5cbe&utm_source=plutonium_analysis_skill&utm_medium=claude_skill&utm_campaign=connector_risk_assessment","privacy_policy_url":"https://agentisend.com/privacy","provisional_risk":null,"publication_recommendation":"fix","publisher":"Agentisend FZCO","publisher_url":"https://agentisend.com/","remote_url":"https://api.agentisend.com/mcp","repository_url":"","risk":"high","risk_severity":"high","security_risks":[{"description":"This connector can publish or send messages and content using the connected identity. Review findings: These tools send external messages as the authenticated sender.","evidence":"Sends, replies to and batches transactional email: The inventory explicitly exposes single, reply and batch email sends, with the listing allowing batches of up to 500 recipients.","manual_risk_type":"messaging_as_user","remediation":{"block_tool_categories":["sends_externally"],"steps":["Require review before messages, posts, or publications are sent.","Restrict delivery to test or internal recipients where possible."]},"risk_type":"sends_messages_as_you","severity":"high","title":"Sends, replies to and batches transactional email","triggering_tools":["reply_to_email","send_batch","send_email"]},{"description":"This connector can permanently remove data or resources. Review findings: These actions destroy persistent customer and integration resources.","evidence":"Deletes contacts and webhooks: The inventory includes explicit deletion of contacts and webhooks.","manual_risk_type":"data_deletion","remediation":{"block_tool_categories":["destructive"],"steps":["Disable destructive tools unless deletion is an explicit workflow requirement.","Require human confirmation and verified backups before deletion."]},"risk_type":"deletes_data","severity":"high","title":"Deletes contacts and webhooks","triggering_tools":["delete_contact","delete_webhook"]},{"description":"This connector can create, update, configure, or otherwise change connected resources. Review findings: OAuth revocation and sender-control changes are security and administrative actions.","evidence":"Revokes OAuth grants and changes sender limits: The connector can revoke an OAuth grant and set a sender limit; its description also identifies budget and kill-switch controls.","manual_risk_type":"credential_admin","remediation":{"block_tool_categories":["state_change"],"steps":["Use a test account or sandbox before allowing changes to production data.","Grant only the write permissions required for the intended workflow.","Require human confirmation for consequential state changes."]},"risk_type":"modifies_data","severity":"high","title":"Revokes OAuth grants and changes sender limits","triggering_tools":["revoke_oauth_grant","set_limit","get_kill_switch_state"]},{"description":"This connector can execute code, commands, queries, or dynamically selected actions. Review findings: The reachable action set is incomplete and can hide additional High operations.","evidence":"Dynamically loads 38 undeclared tools: The listing states that the hosted server exposes 81 tools, only 43 through tools/list and the rest through list_more_tools; the packet inventories only those 43.","manual_risk_type":"generic_gateway","remediation":{"block_tool_categories":["code_execution"],"steps":["Run only in a sandbox with least-privilege credentials and strict resource limits.","Disable generic execution gateways unless their reachable action set is fully understood.","Require approval before execution against production systems or sensitive data."]},"risk_type":"runs_code","severity":"high","title":"Dynamically loads 38 undeclared tools","triggering_tools":["list_more_tools"]}],"signature_status":"unknown","source_code_reviewed":false,"support_url":"hello@agentisend.com","tags":["accesses_calendar","deletes_data","modifies_data","network_access","oauth_required","runs_code","sends_messages_as_you","web_connector"],"tools":[{"description":"","name":"add_suppression"},{"description":"","name":"cancel_email"},{"description":"","name":"create_domain"},{"description":"","name":"create_webhook"},{"description":"","name":"delete_contact","risk":{"category":"destructive","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can permanently remove data or resources. Review findings: These actions destroy persistent customer and integration resources."}},{"description":"","name":"delete_webhook","risk":{"category":"destructive","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can permanently remove data or resources. Review findings: These actions destroy persistent customer and integration resources."}},{"description":"","name":"explain_email"},{"description":"","name":"get_agent_budget"},{"description":"","name":"get_domain"},{"description":"","name":"get_email"},{"description":"","name":"get_email_events"},{"description":"","name":"get_kill_switch_state","risk":{"category":"state_change","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can create, update, configure, or otherwise change connected resources. Review findings: OAuth revocation and sender-control changes are security and administrative actions."}},{"description":"","name":"get_received_email"},{"description":"","name":"get_trust_standing"},{"description":"","name":"list_agent_actions"},{"description":"","name":"list_contacts"},{"description":"","name":"list_domains"},{"description":"","name":"list_emails"},{"description":"","name":"list_more_tools","risk":{"category":"code_execution","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can execute code, commands, queries, or dynamically selected actions. Review findings: The reachable action set is incomplete and can hide additional High operations."}},{"description":"","name":"list_oauth_grants"},{"description":"","name":"list_pending_approvals"},{"description":"","name":"list_received_emails"},{"description":"","name":"list_segments"},{"description":"","name":"list_suppressions"},{"description":"","name":"list_templates"},{"description":"","name":"list_threads"},{"description":"","name":"list_webhooks"},{"description":"","name":"preflight_email"},{"description":"","name":"publish_template"},{"description":"","name":"remove_suppression"},{"description":"","name":"render_template"},{"description":"","name":"reply_to_email","risk":{"category":"sends_externally","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can publish or send messages and content using the connected identity. Review findings: These tools send external messages as the authenticated sender."}},{"description":"","name":"request_approval"},{"description":"","name":"reschedule_email"},{"description":"","name":"revoke_oauth_grant","risk":{"category":"state_change","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can create, update, configure, or otherwise change connected resources. Review findings: OAuth revocation and sender-control changes are security and administrative actions."}},{"description":"","name":"send_batch","risk":{"category":"sends_externally","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can publish or send messages and content using the connected identity. Review findings: These tools send external messages as the authenticated sender."}},{"description":"","name":"send_email","risk":{"category":"sends_externally","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can publish or send messages and content using the connected identity. Review findings: These tools send external messages as the authenticated sender."}},{"description":"","name":"set_limit","risk":{"category":"state_change","level":"high","recommendation":"Follow the connector-level remediation and require approval for consequential actions.","why":"This connector can create, update, configure, or otherwise change connected resources. Review findings: OAuth revocation and sender-control changes are security and administrative actions."}},{"description":"","name":"upsert_contact"},{"description":"","name":"verify_domain"},{"description":"","name":"wait_for_domain"},{"description":"","name":"whoami"},{"description":"","name":"why_was_this_not_sent"}],"tools_count":43,"transport":"streamable-http","type":"web_connector","url":"https://claude.ai/directory/0fbf0daf-cdd3-47f9-8093-20f15f0d5cbe","uuid":"f5ea7ee4-be92-50b3-8b26-33e6d95d43d8","verification_tier":"community","version":""},"query_key":"agentisend fzco agentisend","schema_version":1,"status":"match"}