{"generated_at":"2026-10-14T20:07:22+00:00","item":{"added_at":"2026-07-24","analysis_method":"independent_manual_review_2026-10-10","assessment_status":"provisional","auth_posture":"auth_required","category":"Web Connector","confidence":"medium","description":"Builds and runs AI-agent evaluations in your Coval workspace. Inspect agents, test sets, cases, personas and metrics, or update evaluation fixtures. Launch chat and SMS tests; this connector does not start voice or WebSocket voice runs.","did":"2c91821e-e1f8-481c-83e3-938cc6c6acc4","documentation_url":"https://docs.coval.ai/sofia/overview","editorial_description":{"extended_description":"Create agents, test sets and cases, inspect evaluation-run results, and ask Sofia for read-only analysis or recommendations. The voice restriction also covers outbound voice evaluations.","schema_version":1,"summary":"Builds and runs AI-agent evaluations in your Coval workspace.\nInspect agents, test sets, cases, personas and metrics, or update evaluation fixtures.\nLaunch chat and SMS tests; this connector does not start voice or WebSocket voice runs."},"homepage_url":"","icon":"","id":"2c91821e-e1f8-481c-83e3-938cc6c6acc4","installs":0,"last_scanned":"","license":"","long_description":"Create agents, test sets and cases, inspect evaluation-run results, and ask Sofia for read-only analysis or recommendations. The voice restriction also covers outbound voice evaluations.","manual_assessment":"medium","manual_review":{"all_findings":[{"evidence":"The complete inventory contains persistent create/update operations and create_run.","reasoning":"These tools alter the connected evaluation workspace and launch evaluations.","risk_type":"data_modification","severity":"medium","title":"Creates and updates evaluation resources and launches runs","triggering_tools_or_skills":["create_agent","create_run","create_test_case","create_test_set","update_agent","update_test_case"]}],"assessment_status":"provisional","auth_and_scopes":"Authentication is required for the connected Coval workspace; exact scopes are not documented in the packet.","blocked_missing_evidence":null,"confidence":"medium","displayed_risk":"medium","evidence_completeness":"partial","manual_overall_risk":"medium","official_sources":[{"accessed_at":"2026-10-10T20:24:37Z","url_or_path":"plutosecurity/plutonium@b749c8dc43536c794409be9805edd48063f045e2:app/data/conn/claudesec/2c91821e-e1f8-481c-83e3-938cc6c6acc4.json","what_it_proves":"Candidate PR #125 record identity (name \"Coval\", publisher, type web_connector) and 19 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/2c91821e-e1f8-481c-83e3-938cc6c6acc4: complete tool_names inventory (19 tools), remote streamable-http endpoint https://mcp.coval.dev/claude/mcp, auth_posture auth_required, directory tier community"}],"primary_reason":"Coval modifies evaluation data and starts chat or SMS evaluation runs.","provisional_risk":null,"publication_recommendation":"accept","requires_runtime_verification":true,"review_key":"claude:2c91821e-e1f8-481c-83e3-938cc6c6acc4","reviewed_at":"2026-10-10","reviewer_notes":"The description excludes voice and outbound-voice runs. Runtime verification should establish whether SMS evaluations send real external messages or only simulated test traffic.","risk_domains":{"arbitrary_network_access":"false","authz_ambiguity":"true","browser_control":"false","credential_or_admin_actions":"false","deletes_data":"false","deployment_or_infrastructure":"false","dynamic_hidden_actions":"false","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":"false","sends_messages_as_user":"unknown","spends_or_moves_money":"false"},"tools_or_skills_reviewed":19,"tools_or_skills_total":19},"name":"Coval","needs_runtime_verification":true,"plutonium_url":"https://plutonium.pluto.security/detail.html?planet=claudesec&did=2c91821e-e1f8-481c-83e3-938cc6c6acc4&utm_source=plutonium_analysis_skill&utm_medium=claude_skill&utm_campaign=connector_risk_assessment","privacy_policy_url":"https://www.coval.ai/privacy-policy","provisional_risk":null,"publication_recommendation":"accept","publisher":"Coval","publisher_url":"https://www.coval.ai/","remote_url":"https://mcp.coval.dev/claude/mcp","repository_url":"","risk":"medium","risk_severity":"medium","security_risks":[{"description":"This connector can create, update, configure, or otherwise change connected resources. Review findings: These tools alter the connected evaluation workspace and launch evaluations.","evidence":"Creates and updates evaluation resources and launches runs: The complete inventory contains persistent create/update operations and create_run.","manual_risk_type":"data_modification","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":"medium","title":"Creates and updates evaluation resources and launches runs","triggering_tools":["create_agent","create_run","create_test_case","create_test_set","update_agent","update_test_case"]}],"signature_status":"unknown","source_code_reviewed":false,"support_url":"support@coval.dev","tags":["modifies_data","network_access","oauth_required","web_connector"],"tools":[{"description":"","name":"consult_sofia"},{"description":"","name":"create_agent","risk":{"category":"state_change","level":"medium","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: These tools alter the connected evaluation workspace and launch evaluations."}},{"description":"","name":"create_run","risk":{"category":"state_change","level":"medium","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: These tools alter the connected evaluation workspace and launch evaluations."}},{"description":"","name":"create_test_case","risk":{"category":"state_change","level":"medium","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: These tools alter the connected evaluation workspace and launch evaluations."}},{"description":"","name":"create_test_set","risk":{"category":"state_change","level":"medium","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: These tools alter the connected evaluation workspace and launch evaluations."}},{"description":"","name":"get_agent"},{"description":"","name":"get_metric"},{"description":"","name":"get_persona"},{"description":"","name":"get_run"},{"description":"","name":"get_test_case"},{"description":"","name":"get_test_set"},{"description":"","name":"list_agents"},{"description":"","name":"list_metrics"},{"description":"","name":"list_personas"},{"description":"","name":"list_runs"},{"description":"","name":"list_test_cases"},{"description":"","name":"list_test_sets"},{"description":"","name":"update_agent","risk":{"category":"state_change","level":"medium","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: These tools alter the connected evaluation workspace and launch evaluations."}},{"description":"","name":"update_test_case","risk":{"category":"state_change","level":"medium","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: These tools alter the connected evaluation workspace and launch evaluations."}}],"tools_count":19,"transport":"streamable-http","type":"web_connector","url":"https://claude.ai/directory/2c91821e-e1f8-481c-83e3-938cc6c6acc4","uuid":"1c04cbc2-0fad-55d5-a3a5-74941b1beaae","verification_tier":"community","version":""},"query_key":"coval coval","schema_version":1,"status":"match"}