Countly MCP Policy -- 127 Tools

// TOOL MAP

Read (69) Write / Execute (38) Destructive / Financial (20)

// WHAT CAN GO WRONG

✕

Destructive tools (ab_experiments_delete, alerts_delete, app_users_delete) permanently delete resources. There is no undo. An agent calling these in a retry loop causes irreversible damage.

✕

Write operations (ab_experiments_create, alerts_create, app_users_create) modify state. Without rate limits, an agent can make hundreds of changes in seconds — faster than any human can review or revert.

✕

Execute tools (ab_experiments_start, ab_experiments_stop, collections_aggregate) trigger processes with side effects. Builds, notifications, workflows — all fired without throttling.

// RECOMMENDED RULES

Deny destructive operations

ab_experiments_delete:
  rules:
    - action: deny

Destructive tools should never be available to autonomous agents without human approval.

Rate limit write operations

ab_experiments_create:
  rules:
    - rate_limit: 30/hour

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

ab_experiments_details:
  rules:
    - rate_limit: 60/minute

Controls API costs and prevents retry loops from exhausting upstream rate limits.

// ALL 127 COUNTLY TOOLS

EXECUTE 4 tools

Execute ab_experiments_start Execute ab_experiments_stop Execute collections_aggregate Execute formulas_run

WRITE 34 tools

Write ab_experiments_create Write alerts_create Write app_users_create Write app_users_update Write apps_create Write apps_update Write cohorts_create Write cohorts_update Write crashes_comment_add Write crashes_comment_update Write crashes_resolve Write dashboards_create Write dashboards_update Write dashboards_widget_add Write dashboards_widget_update Write drill_bookmarks_create Write email_reports_core_create Write email_reports_dashboard_create Write email_reports_preview Write email_reports_send Write email_reports_update Write events_create Write filtering_rules_create Write filtering_rules_update Write formulas_save Write funnels_create Write funnels_update Write hooks_create Write hooks_update Write notes_create Write remote_config_conditions_add Write remote_config_conditions_update Write remote_config_parameters_add Write remote_config_parameters_update

READ 69 tools

Read ab_experiments_details Read ab_experiments_list Read alerts_list Read app_analytics_summary Read apps_get_by_name Read apps_list Read cohorts_data Read cohorts_list Read collections_indexes Read consents_history_search Read consents_list Read consents_stats Read crash_groups_list Read crashes_get Read crashes_hide Read crashes_show Read crashes_stats_get Read dashboard_users Read dashboards_data Read dashboards_list Read databases_document Read databases_list Read databases_query Read databases_stats Read datapoints_punch_card Read datapoints_stats Read datapoints_top_apps Read drill_bookmarks_list Read email_reports_list Read events_list Read filtering_rules_list Read filtering_rules_toggle_status Read formulas_list Read funnels_data Read funnels_dropoff_users Read funnels_list Read funnels_step_users Read get_plugins Read get_version Read hooks_list Read hooks_test Read live_last_30_days Read live_last_day Read live_last_hour Read live_metrics Read live_overall Read live_users Read metadata_get Read notes_list Read ping Read queriable_fields_list Read query_data Read remote_configs_list Read retention Read sdk_config_get Read sdk_logs_list Read sdk_stats_get Read server_logs_contents Read server_logs_files_list Read session_durations Read session_frequency Read slipping_users Read times_of_day Read user_loyalty Read user_profiles_breakdown Read user_profiles_get Read user_profiles_query Read views_data Read views_table

// FAQ

Can an AI agent delete data through the Countly MCP server? +

Yes. The Countly server exposes 20 destructive tools including ab_experiments_delete, alerts_delete, app_users_delete. These permanently remove resources with no undo. Intercept blocks destructive tools by default so they never reach the upstream server.

How do I prevent bulk modifications through Countly? +

The Countly server has 34 write tools including ab_experiments_create, alerts_create, app_users_create. Set rate limits in your policy file -- for example, rate_limit: 10/hour prevents an agent from making more than 10 modifications per hour. Intercept enforces this at the transport layer.

How many tools does the Countly MCP server expose? +

127 tools across 4 categories: Destructive, Execute, Read, Write. 69 are read-only. 58 can modify, create, or delete data.

How do I add Intercept to my Countly setup? +

One line change. Instead of running the Countly server directly, prefix it with Intercept: intercept -c countly.yaml -- npx -y @countly-mcp-server. Download a pre-built policy from policylayer.com/policies/countly and adjust the limits to match your use case.

// RELATED SERVERS

Other MCP servers with similar tools.

Starter policies available for each. Same risk classification, same one-command setup.

Countly

Other MCP servers with similar tools.

Let agents act without letting them run wild.