@thinkingai/ae-cli 6.0.30 → 6.0.32

package/skills/ae-analysis/references/create_project_mark_time.md

@@ -7,6 +7,7 @@ Domain: **Project Configuration**
 ## Use Cases
 - Create a project date marker. Returns the new marker ID, timestamp, content, and visibility status.
 - Create a project date marker.
+- Do not use it to modify an existing marker; use `+update_project_mark_time` with the verified marker ID.
 
 ## Command
 ```bash
@@ -28,7 +29,7 @@ ae-cli analysis_meta +create_project_mark_time --dry-run
 - First run should only pass required parameters (`--project_id`, `--marked_at`, `--content`), and add optional parameters only after the path is confirmed to work.
 - For date/time ranges, first validate with a short range, then expand the range step by step.
 - For cross-project troubleshooting, first confirm whether `--project_id` matches the current permissions and target environment.
-- Write operations keep the confirmation prompt by default; re-evaluate whether to use `--yes` in automation scenarios.
+- This is an ordinary `write` operation and does not require CLI confirmation.
 
 ## Next Steps After Failure
 - If required parameters are missing, fall back to the smallest runnable command and fill them in (focus on `--project_id`, `--marked_at`, and `--content`).

package/skills/ae-analysis/references/dashboard_copy.md

@@ -9,7 +9,7 @@ If both `--to-space-id` and `--to-folder-id` are omitted, the gateway copies to
 Command:
 
 ```bash
-ae-cli analysis dashboard copy --project-id <project_id> --dashboard-id <source_dashboard_id> --dashboard-name <new_name> [--report-copy true] [--to-space-id <space_id>] [--to-folder-id <folder_id>] --yes
+ae-cli analysis dashboard copy --project-id <project_id> --dashboard-id <source_dashboard_id> --dashboard-name <new_name> [--report-copy true] [--to-space-id <space_id>] [--to-folder-id <folder_id>]
 ```
 
 Input sends `project_id`, `dashboard_id`, `dashboard_name`, `report_copy` defaulting to false, and optional `to_space_id`, `to_folder_id`.

package/skills/ae-analysis/references/dashboard_create.md

@@ -7,7 +7,7 @@ Do not use to copy an existing dashboard. Use `dashboard copy` for copy workflow
 Command:
 
 ```bash
-ae-cli analysis dashboard create --project-id <project_id> --dashboard-name <name> [--space-id <space_id>] [--folder-id <folder_id>] [--initial-report-id <report_id>] [--payload '{...}'] --yes
+ae-cli analysis dashboard create --project-id <project_id> --dashboard-name <name> [--space-id <space_id>] [--folder-id <folder_id>] [--initial-report-id <report_id>] [--payload '{...}']
 ```
 
 Input sends `project_id`, `dashboard_name`, and optional `space_id`, `folder_id`, `initial_report_id`, `payload`.

package/skills/ae-analysis/references/dashboard_daily_report_send.md

@@ -7,7 +7,7 @@ Do not use for configuration changes. Use `dashboard-daily-report update`.
 Command:
 
 ```bash
-ae-cli analysis dashboard-daily-report send --project-id <project_id> --dashboard-id <dashboard_id> [--need-csv true] [--host-url <url>] [--payload '{...}'] --yes
+ae-cli analysis dashboard-daily-report send --project-id <project_id> --dashboard-id <dashboard_id> [--need-csv true] [--host-url <url>] [--payload '{...}']
 ```
 
 Input sends `project_id`, `dashboard_id`, and optional daily report fields or `payload`.

package/skills/ae-analysis/references/dashboard_daily_report_update.md

@@ -7,7 +7,7 @@ Do not use to send immediately. Use `dashboard-daily-report send`.
 Command:
 
 ```bash
-ae-cli analysis dashboard-daily-report update --project-id <project_id> --dashboard-id <dashboard_id> [--enable-send true] [--send-time <time>] [--send-title <title>] [--send-content <content>] [--payload '{...}'] --yes
+ae-cli analysis dashboard-daily-report update --project-id <project_id> --dashboard-id <dashboard_id> [--enable-send true] [--send-time <time>] [--send-title <title>] [--send-content <content>] [--payload '{...}']
 ```
 
 Input sends `project_id`, `dashboard_id`, and optional daily report fields or `payload`.

package/skills/ae-analysis/references/dashboard_definition_import.md

@@ -7,7 +7,7 @@ Do not use a separate pre-check command; validation is part of this import comma
 Command:
 
 ```bash
-ae-cli analysis dashboard-definition import --project-id <project_id> --definition '{...}' [--validate-only true] [--dashboard-name-conflict-policy <policy>] [--space-dashboard-policy <policy>] [--payload '{...}'] --yes
+ae-cli analysis dashboard-definition import --project-id <project_id> --definition '{...}' [--validate-only true] [--dashboard-name-conflict-policy <policy>] [--space-dashboard-policy <policy>] [--payload '{...}']
 ```
 
 Input sends `project_id`, `definition`, and optional validation/import policies and `payload`.

package/skills/ae-analysis/references/dashboard_delete.md

@@ -7,6 +7,8 @@ Do not use for soft archival or hiding unless the gateway contract defines that
 Command:
 
 ```bash
+ae-cli analysis dashboard delete --project-id <project_id> --dashboard-ids '[1001,1002]' --dry-run
+# Summarize the target and impact, then wait for explicit user confirmation.
 ae-cli analysis dashboard delete --project-id <project_id> --dashboard-ids '[1001,1002]' --yes
 ```
 

package/skills/ae-analysis/references/dashboard_freeze.md

@@ -7,7 +7,7 @@ Do not use for dashboard lock/version operations; those are intentionally not CL
 Command:
 
 ```bash
-ae-cli analysis dashboard freeze --project-id <project_id> --dashboard-ids '[1001,1002]' [--freeze true] --yes
+ae-cli analysis dashboard freeze --project-id <project_id> --dashboard-ids '[1001,1002]' [--freeze true]
 ```
 
 Input sends `project_id`, `dashboard_ids`, and optional `freeze`.

package/skills/ae-analysis/references/dashboard_handover.md

@@ -7,7 +7,7 @@ Do not use to share dashboards without ownership transfer. Use `dashboard share`
 Command:
 
 ```bash
-ae-cli analysis dashboard handover --project-id <project_id> --dashboard-ids '[1001,1002]' --to-user-id <user_id> --yes
+ae-cli analysis dashboard handover --project-id <project_id> --dashboard-ids '[1001,1002]' --to-user-id <user_id>
 ```
 
 Input sends `project_id`, `dashboard_ids`, and `to_user_id`.

package/skills/ae-analysis/references/dashboard_list.md

@@ -7,9 +7,11 @@ Do not use for dashboard report data. Use `dashboard-report-data run` or `dashbo
 Command:
 
 ```bash
-ae-cli analysis dashboard list --project-id <project_id> [--query <keyword>] [--fields '["id","name"]'] [--limit 20] [--offset 0]
+ae-cli analysis dashboard list --project-id <project_id> [--query <keyword>] [--fields '["dashboard_id","dashboard_name"]'] [--limit 50] [--offset 0]
 ```
 
-Input uses `project_id` plus optional `query`, `fields`, `limit`, `offset`. Use `--payload` only when the dedicated gateway contract requires extra snake_case filters.
+Input uses `project_id` plus optional `query`, `fields`, `limit`, `offset`. `--fields` accepts only `dashboard_id`, `dashboard_name`, and `remark`. Input and output both use snake_case; do not use `dashboardId`, `dashboardName`, generic `id`, or generic `name`.
+
+When `has_more=true`, continue only with the returned `next_offset`; do not calculate the next page locally.
 
 Output is the gateway envelope. `data` contains dashboard summaries and paging metadata when returned by the gateway.

package/skills/ae-analysis/references/dashboard_report_add.md

@@ -0,0 +1,15 @@
+# analysis dashboard-report add
+
+Use when the user explicitly wants to add existing saved reports to a dashboard.
+
+Do not use for creating reports or dashboards. Create those assets first, then call this command with exact IDs.
+
+Command:
+
+```bash
+ae-cli analysis dashboard-report add --project-id <project_id> --dashboard-id <dashboard_id> --report-ids '[1001,1002]'
+```
+
+Input sends `project_id`, `dashboard_id`, and `report_ids`.
+
+Output is the gateway envelope. `data` contains `dashboard_id`, requested `report_ids`, and `added_count`.

package/skills/ae-analysis/references/dashboard_report_data_export.md

@@ -1,8 +1,10 @@
 # analysis dashboard-report-data export
 
-Use for dashboard report data queries that may be large or long-running.
+Submit dashboard report data as a downloadable async artifact. Dashboard report data follows report-data model coverage: the 12 analysis report models from `ai_models.md` plus tag report data.
 
-Do not use when the user needs a tiny immediate preview. Use `dashboard-report-data run` with a small `--limit`.
+Routing: read [`analysis_data_retrieval.md`](analysis_data_retrieval.md) before choosing this `export` command instead of `dashboard-report-data run`.
+
+Do not use this command for bounded inline previews; use `dashboard-report-data run` when the requested result fits the sync data retrieval rule.
 
 Command:
 
@@ -10,15 +12,21 @@ Command:
 ae-cli analysis dashboard-report-data export --project-id <project_id> --dashboard-id <dashboard_id> [--report-ids '[1,2]'] [--filters '{...}'] [--start-time <time>] [--end-time <time>] [--artifact-format jsonl]
 ```
 
-Input sends `project_id`, `dashboard_id`, and optional `report_ids`, `filters`, `start_time`, `end_time`, `use_cache`, `request_id`, `timeout_seconds`, `format`. Use CLI flag `--artifact-format` for the gateway `format` input; `--format` is the CLI output formatter.
+Input sends `project_id`, `dashboard_id`, and optional `report_ids`, `filters`, `start_time`, `end_time`, `use_cache`, `request_id`, `timeout_seconds`, `format`. Use CLI flag `--artifact-format` for the gateway `format` input; `--format` is the CLI output formatter. Async export has no inline row limit. Runtime defaults to and is capped at 21600 seconds (6 hours); cancel earlier with `analysis query cancel --run-id <run_id>`. The routing rule lives in [`analysis_data_retrieval.md`](analysis_data_retrieval.md).
 
-`--filters` is an analysis Filter JSON injected as `commonFilter.aiFilter` for each queried report. It is not a dashboard UI control dump. Use `ae-cli analysis +get_filter_schema` for the exact schema. Minimal shape:
+`--filters` uses the same AI-facing filter model as report data. The gateway resolves fields and compiles it to the dashboard backend filter. Do not pass a dashboard UI control dump or raw QP filter. Minimal shape:
 
 ```json
-{"relation":"and","filts":[{"field":{"type":"event_property","propertyName":"#event_name"},"operator":"equal","values":["login"]}]}
+{"relation":"and","items":[{"field":{"name":"country","type":"user_property"},"operator":"eq","values":["US"]}]}
 ```
 
-Output is the gateway envelope. `data` contains an async export descriptor with `run_id`, `artifact_id`, status fields, and expiration fields. It does not expose inspect/download API paths; use the CLI commands below.
+Dashboard `filters`, `start_time`, and `end_time` do not apply to SQL reports. Export still completes with SQL report data; the JSONL result object includes `warnings[]` with `OVERRIDE_IGNORED_FOR_MODEL`, affected SQL `report_ids`, and `ignored_fields`. Query SQL reports directly with report-data `--sql-params` when SQL conditions must change.
+
+Output is the gateway envelope. `data` contains an async export descriptor with `run_id`, `artifact_id`, status fields, and expiration fields. Exports do not create `query_context_id`. It does not expose inspect/download API paths; use the CLI commands below.
+
+An empty artifact is successful and means the requested time range has no data. The run reaches `FAILED` only when every returned dashboard report entry contains an explicit execution error. Mixed dashboard exports retain successful data and explicit per-report failure entries.
+
+Never use the export response or downloaded rows as a drilldown/result-cluster source. Run a bounded synchronous dashboard report preview containing the desired cell first.
 
 Follow-up workflow:
 
@@ -26,6 +34,6 @@ Follow-up workflow:
 2. Poll status with `ae-cli analysis run inspect --run-id <run_id>`.
 3. Continue polling while status is running or pending. Treat `COMPLETED` or `SUCCEEDED` as success, and `FAILED`, `CANCELED`, or `CANCELLED` as terminal failure.
 4. On success, download with `ae-cli analysis artifact download --run-id <run_id> --artifact-id <artifact_id> --output <file>`.
-5. If the export is no longer needed, cancel with `ae-cli analysis query cancel --run-id <run_id> --yes`.
+5. If the export is no longer needed, cancel with `ae-cli analysis query cancel --run-id <run_id>`.
 
 Do not write custom Python/curl for polling or download unless the CLI command itself is unavailable.

package/skills/ae-analysis/references/dashboard_report_data_run.md

@@ -1,21 +1,29 @@
 # analysis dashboard-report-data run
 
-Use for bounded inline dashboard report data queries when the expected result is small enough for immediate JSON output.
+Execute bounded inline dashboard report data queries. Dashboard report data follows report-data model coverage: the 12 analysis report models from `ai_models.md` plus tag report data.
 
-Do not use for large or long-running result sets. Use `dashboard-report-data export`.
+Routing: read [`analysis_data_retrieval.md`](analysis_data_retrieval.md) before choosing this `run` command instead of `dashboard-report-data export`.
+
+Do not use this command for full, unknown-size, larger than 1000-row, or long-running dashboard report data; use `dashboard-report-data export`.
 
 Command:
 
 ```bash
-ae-cli analysis dashboard-report-data run --project-id <project_id> --dashboard-id <dashboard_id> [--report-ids '[1,2]'] [--filters '{...}'] [--start-time <time>] [--end-time <time>] [--limit 100] [--timeout-seconds 60]
+ae-cli analysis dashboard-report-data run --project-id <project_id> --dashboard-id <dashboard_id> [--report-ids '[1,2]'] [--filters '{...}'] [--start-time <time>] [--end-time <time>] [--limit 100] [--timeout-seconds 180]
 ```
 
-Input sends `project_id`, `dashboard_id`, and optional `report_ids`, `filters`, `start_time`, `end_time`, `use_cache`, `request_id`, `timeout_seconds`, `limit`.
+Input sends `project_id`, `dashboard_id`, and optional `report_ids`, `filters`, `start_time`, `end_time`, `use_cache`, `request_id`, `timeout_seconds`, `limit`. Control defaults: `--limit` default 100 / max 1000, `--timeout-seconds` default and max 180. An explicitly supplied lower timeout still wins. The routing rule lives in [`analysis_data_retrieval.md`](analysis_data_retrieval.md).
 
-`--filters` is an analysis Filter JSON injected as `commonFilter.aiFilter` for each queried report. It is not a dashboard UI control dump. Use `ae-cli analysis +get_filter_schema` for the exact schema. Minimal shape:
+`--filters` uses the same AI-facing filter model as report data. The gateway resolves fields and compiles it to the dashboard backend filter. Do not pass a dashboard UI control dump or raw QP filter. Minimal shape:
 
 ```json
-{"relation":"and","filts":[{"field":{"type":"event_property","propertyName":"#event_name"},"operator":"equal","values":["login"]}]}
+{"relation":"and","items":[{"field":{"name":"country","type":"user_property"},"operator":"eq","values":["US"]}]}
 ```
 
-Output is the gateway envelope. `data` contains bounded inline report data.
+Dashboard `filters`, `start_time`, and `end_time` do not apply to SQL reports. The query still succeeds and returns those SQL report results. Inspect `data.warnings[]`; `OVERRIDE_IGNORED_FOR_MODEL` contains the affected SQL `report_ids` and `ignored_fields`. To change SQL conditions, query the SQL report directly with `analysis report-data run` and saved `definition.params` through `--sql-params`.
+
+Output is the gateway envelope. `data` contains bounded inline report data, optional structured `warnings`, plus `query_context_id` and per-report `sources[].drilldown` options when a returned report exposes follow-up actions. Unsupported report models have empty actions.
+
+An empty dashboard batch or report result with no rows is a successful query: it means the requested time range has no data. The command fails only when every returned report entry contains an explicit execution error. Mixed dashboard batches keep successful data and return `meta.partial`, counts, and per-report `meta.failures`.
+
+Read [`analysis_drilldown_contract.md`](analysis_drilldown_contract.md). Pass the selected report source and merge only its returned coordinate fragments for the advertised action. Do not pass raw QP.

package/skills/ae-analysis/references/dashboard_share.md

@@ -7,9 +7,18 @@ Do not use to read sharing only. Use `dashboard share-info`.
 Command:
 
 ```bash
-ae-cli analysis dashboard share --project-id <project_id> --dashboard-id <dashboard_id> [--member-authorities '{...}'] [--payload '{...}'] --yes
+ae-cli analysis dashboard share --project-id <project_id> --dashboard-id <dashboard_id> --member-authorities '{"10001":"READ","10002":"EDIT"}'
 ```
 
-Input sends `project_id`, `dashboard_id`, and either `member_authorities` or a backend-compatible snake_case `payload`.
+Input sends `project_id`, `dashboard_id`, and either `member_authorities` or a documented snake_case `payload`.
+
+`member_authorities` is the complete directly shared user map, not an incremental patch:
+
+- each key is a numeric user ID encoded as a JSON object key;
+- each value is `READ`, `EDIT`, `CREATOR`, or `MAINTAIN`;
+- users omitted from the map are removed from direct sharing;
+- `{}` removes all directly shared users.
+
+Call `analysis dashboard share-info` first when the user intends to preserve existing members while adding or changing only some entries.
 
 Output is the gateway envelope. `data` contains the share update result.

package/skills/ae-analysis/references/dashboard_update.md

@@ -7,10 +7,19 @@ Do not use for moving/copying dashboards. Use `dashboard copy`, `dashboard hando
 Command:
 
 ```bash
-ae-cli analysis dashboard update --project-id <project_id> --operation settings --dashboard-ids '[1001,1002]' [--zone-offset 8] [--payload '{...}'] --yes
-ae-cli analysis dashboard update --project-id <project_id> --operation note-upsert --dashboard-id <dashboard_id> [--note-id <note_id>] [--note-title <title>] [--description <text>] --yes
+ae-cli analysis dashboard update --project-id <project_id> --operation settings --dashboard-ids '[1001,1002]' [--zone-offset 8] [--payload '{...}']
+ae-cli analysis dashboard update --project-id <project_id> --operation settings --dashboard-id <dashboard_id> --refresh-type 1 --dashboard-status normal --payload '{"dashboard_job_schedule":"0 0 8 * * ?","time_config_open":true,"time_config":{},"cache_config":{},"schedule_ui_config":{}}'
+ae-cli analysis dashboard update --project-id <project_id> --operation note-upsert --dashboard-id <dashboard_id> [--note-id <note_id>] [--note-title <title>] [--description <text>]
 ```
 
-Input sends `project_id`, `operation`, optional dashboard identifiers, setting fields (including `zone_offset` as hours, e.g. `8` for UTC+8), note fields, and optional `payload`.
+For `operation=settings`:
+
+- identify one dashboard with `dashboard_id`, or a batch with `dashboard_ids`;
+- `refresh_type` is integer `0` (real-time) or `1` (scheduled);
+- `dashboard_status` is string `normal` or `freeze`;
+- `zone_offset` is an integer hour offset from `-12` to `14`;
+- complex settings belong in snake_case `payload`: `reports_version` string, `dashboard_job_schedule` string, `time_config_open` boolean, and object fields `time_config`, `cache_config`, `schedule_ui_config`; `ui_config` may be a string or object.
+
+For `operation=note-upsert`, pass `dashboard_id`; omit `note_id` to create a note or pass it to update an existing note. Do not mix note fields with batch settings fields.
 
 Output is the gateway envelope. `data` contains the update result.

package/skills/ae-analysis/references/datatable_columns_get.md

@@ -2,7 +2,7 @@
 
 Use when the user needs to get columns for a project table reference.
 
-Do not use this command for unrelated analysis queries, ad-hoc query construction, or MCP metadata discovery when an existing specialized command already fits the user's request.
+Do not use it to read table rows or infer a table reference; discover the exact table first and use a data query for row values.
 
 Command:
 
@@ -15,7 +15,7 @@ Capability id: `metadata.data_table.columns_get`.
 
 Input sends `project_id`, `table_ref`.
 
-Output is the gateway envelope. `data` contains the common-service capability result.
+Output contains resolved `data.table_ref` and `data.columns[]` from SQL IDE metadata.
 
 ## Parameters
 | Parameter | Required | Description |

package/skills/ae-analysis/references/datatable_influence_list.md

@@ -2,7 +2,7 @@
 
 Use when the user needs to list metadata and assets affected by data table deletion or column changes.
 
-Do not use this command for unrelated analysis queries, ad-hoc query construction, or MCP metadata discovery when an existing specialized command already fits the user's request.
+Do not use it as a delete command. It is a read-only dependency check before table/column mutation.
 
 Command:
 
@@ -15,7 +15,7 @@ Capability id: `metadata.data_table.influence_list`.
 
 Input sends `project_id`, `datatable_id`.
 
-Output is the gateway envelope. `data` contains the common-service capability result.
+Output includes `data.resources[]`, `data.continuable`, and `data.project_info` for the proposed table operation.
 
 ## Parameters
 | Parameter | Required | Description |

package/skills/ae-analysis/references/datatable_version_get.md

@@ -2,7 +2,7 @@
 
 Use when the user needs to get data table version detail.
 
-Do not use this command for unrelated analysis queries, ad-hoc query construction, or MCP metadata discovery when an existing specialized command already fits the user's request.
+Do not use it for the current table definition without a verified version ID; list versions first.
 
 Command:
 
@@ -15,7 +15,7 @@ Capability id: `metadata.data_table_version.get`.
 
 Input sends `project_id`, `version_id`.
 
-Output is the gateway envelope. `data` contains the common-service capability result.
+Output `data.version` contains the selected historical version detail.
 
 ## Parameters
 | Parameter | Required | Description |

package/skills/ae-analysis/references/datatable_version_list.md

@@ -2,7 +2,7 @@
 
 Use when the user needs to list data table historical versions.
 
-Do not use this command for unrelated analysis queries, ad-hoc query construction, or MCP metadata discovery when an existing specialized command already fits the user's request.
+Do not use it for one version's full definition or table row history; use `datatable version-get` for the selected metadata version.
 
 Command:
 
@@ -15,7 +15,7 @@ Capability id: `metadata.data_table_version.list`.
 
 Input sends `project_id`, `datatable_id`.
 
-Output is the gateway envelope. `data` contains the common-service capability result.
+Output `data.versions[]` contains historical metadata versions for the data table.
 
 ## Parameters
 | Parameter | Required | Description |

package/skills/ae-analysis/references/delete_alert.md

@@ -6,11 +6,16 @@ Domain: **Alert Management**
 
 ## Use Cases
 - Permanently delete an alert by its ID.
+- Do not use this command to disable or edit an alert; use `+update_alert` when the alert must remain.
+
+## Output
+Success confirms that the specified alert was deleted. The command does not return alert data; use `+list_alerts` to verify it is absent.
 
 ## Commands
 ```bash
-ae-cli analysis +delete_alert --project_id <project_id> --alert_id <alert_id>
-ae-cli analysis +delete_alert --dry-run
+ae-cli analysis +delete_alert --project_id <project_id> --alert_id <alert_id> --dry-run
+# Summarize the target and impact, then wait for explicit user confirmation.
+ae-cli analysis +delete_alert --project_id <project_id> --alert_id <alert_id> --yes
 ```
 
 ## Parameters
@@ -21,7 +26,7 @@ ae-cli analysis +delete_alert --dry-run
 
 ## Decision Rules
 - Use `+list_alerts` first to confirm the alert ID before deleting.
-- This is a destructive operation; keep the confirmation prompt unless automation is explicitly required.
+- This is `high-risk-write`: inspect the dry-run, summarize the target and impact, and wait for explicit user confirmation before the `--yes` execution.
 
 ## Recommended Chain
 - `+list_alerts` -> `+delete_alert`

package/skills/ae-analysis/references/delete_project_mark_times.md

@@ -6,11 +6,13 @@ Domain: **Project Configuration**
 
 ## Use Cases
 - Delete one or more project date markers.
+- Do not use it to hide or edit a marker; use `+update_project_mark_time` when the marker should remain.
 
 ## Commands
 ```bash
-ae-cli analysis_meta +delete_project_mark_times --project_id <project_id> --mark_time_ids '[]'
-ae-cli analysis_meta +delete_project_mark_times --dry-run
+ae-cli analysis_meta +delete_project_mark_times --project_id <project_id> --mark_time_ids '[]' --dry-run
+# Summarize the target and impact, then wait for explicit user confirmation.
+ae-cli analysis_meta +delete_project_mark_times --project_id <project_id> --mark_time_ids '[]' --yes
 ```
 
 ## Parameters
@@ -23,7 +25,7 @@ ae-cli analysis_meta +delete_project_mark_times --dry-run
 - For the first run, pass only the required parameters (`--project_id` and `--mark_time_ids`) to confirm the path works, then add optional parameters.
 - Wrap JSON parameters in single quotes (for example `--mark_time_ids '{}'`) to avoid shell escaping issues.
 - For cross-project troubleshooting, first confirm whether `--project_id` matches the current permissions and target environment.
-- Write operations keep confirmation prompts by default; evaluate whether to use `--yes` only for automation scenarios.
+- This is `high-risk-write`: inspect the dry-run, summarize the target and impact, and wait for explicit user confirmation before the `--yes` execution.
 
 ## Next Steps After Failure
 - If required parameters are missing, fall back to the smallest runnable command and fill them in first (focus on `--project_id` and `--mark_time_ids`).

package/skills/ae-analysis/references/delete_track_items.md

@@ -10,8 +10,9 @@ Domain: **Project Configuration**
 
 ## Command
 ```bash
-ae-cli analysis_meta +delete_track_items --project_id <project_id> --delete_data '{}'
-ae-cli analysis_meta +delete_track_items --dry-run
+ae-cli analysis_meta +delete_track_items --project_id <project_id> --delete_data '{}' --dry-run
+# Summarize the target and impact, then wait for explicit user confirmation.
+ae-cli analysis_meta +delete_track_items --project_id <project_id> --delete_data '{}' --yes
 ```
 
 ## Parameters
@@ -24,7 +25,7 @@ ae-cli analysis_meta +delete_track_items --dry-run
 - First run should only pass required parameters (`--project_id`, `--delete_data`), and add optional parameters only after the path is confirmed to work.
 - Wrap JSON arguments in single quotes (for example `--delete_data '{}'`) to avoid shell escaping issues.
 - For cross-project troubleshooting, first confirm whether `--project_id` matches the current permissions and target environment.
-- Write operations keep the confirmation prompt by default; re-evaluate whether to use `--yes` in automation scenarios.
+- This is `high-risk-write`: inspect the dry-run, summarize the target and impact, and wait for explicit user confirmation before the `--yes` execution.
 
 ## Next Steps After Failure
 - If required parameters are missing, fall back to the smallest runnable command and fill them in (focus on `--project_id` and `--delete_data`).

package/skills/ae-analysis/references/drilldown_entities_export.md

@@ -0,0 +1,26 @@
+# analysis drilldown-entities export
+
+Stream the complete users or custom entities behind one selected synchronous-preview coordinate as a `csv.gz` artifact.
+
+Read [`analysis_drilldown_contract.md`](analysis_drilldown_contract.md) and first validate the coordinate with the corresponding synchronous preview/`drilldown-entities run` path. The export consumes an existing sync-preview context; it never creates a query context or additional drilldown choices.
+
+```bash
+ae-cli analysis drilldown-entities export \
+  --query-context-id <sync_preview_query_context_id> \
+  [--source '{"report_id":1001}'] \
+  --coordinate '<same returned coordinate>' \
+  [--properties '[{...}]'] \
+  [--artifact-format csv] \
+  [--timeout-seconds 21600]
+```
+
+Property support matches the synchronous preview exactly:
+
+- For the `#user_id` user subject, `--properties` accepts `[{"columnName":"<property_name>","tableType":"user"}]`. The artifact always retains `#user_id`, `#account_id`, and `#distinct_id` plus the requested user properties.
+- For a custom entity subject, omit `--properties`; the artifact contains only the entity value column. Supplying properties returns `CUSTOM_ENTITY_PROPERTIES_UNSUPPORTED` instead of silently ignoring them.
+
+`#user_id` is an internal association key. When presenting downloaded user rows to a customer, Agents should display account ID and visitor ID by default rather than using `#user_id` as the only visible identity.
+
+This command does not accept `--limit`, `--offset`, `--page-num`, or `--page-size`. Common executes one full-download query and streams it directly into the artifact; it does not repeatedly call the synchronous preview. “Complete” follows the platform full-download ceiling (`model_full_download_limit`). Inspect `run_id`, then download only after completion.
+
+The file is durable member data, not a new interactive result. Do not use rows in the downloaded artifact to construct another analysis coordinate. User event continuation requires the `drilldown_context_id` and canonical `user_id` returned by a user-subject `drilldown-entities run`, not this export.

package/skills/ae-analysis/references/drilldown_entities_run.md

@@ -0,0 +1,33 @@
+# analysis drilldown-entities run
+
+Preview users or custom entities behind one selected synchronous-preview cell.
+
+Read [`analysis_drilldown_contract.md`](analysis_drilldown_contract.md) first. Call this only when the selected metric/action advertises `drilldown_entities`. Event metrics require `analysis_angle=USER_LIST|ENTITY_LIST`; retention, funnel, distribution, interval, property, path, and attribution results expose their analysis subject in `source.drilldown.subject`.
+
+## Command
+
+```bash
+ae-cli analysis drilldown-entities run \
+  --query-context-id <sync_preview_query_context_id> \
+  [--source '{"report_id":1001}'] \
+  --coordinate '<merged returned row/column/metric coordinate>' \
+  [--properties '[{...}]'] \
+  [--limit 100] \
+  [--timeout-seconds 120]
+```
+
+The context, source, and coordinate must all come from the same synchronous preview. Select a visible row from `row_options`, a drillable column from `column_options`, and the relevant metric option; shallow-merge their coordinate fragments. Never send `row_index`, `column_index`, `values`, `label`, `target_id`, raw QP, or a coordinate derived from an export file.
+
+Property support depends on the returned subject:
+
+- `subject.type=user` (`column_name=#user_id`): `--properties` may select additional user properties with `[{"columnName":"<property_name>","tableType":"user"}]`. The named value `user` is required; do not send numeric enum codes. Every row still contains `#user_id`, `#account_id`, and `#distinct_id`.
+- `subject.type=entity`: do not pass `--properties`. A custom entity returns only its entity value. Supplying properties is invalid and must return `CUSTOM_ENTITY_PROPERTIES_UNSUPPORTED`; it is never silently ignored.
+
+## Output and next intent
+
+The response contains `subject`, `items`, `total`, `returned_rows`, and `truncated`.
+
+- `subject.type=user`: items include canonical `user_id`; `attributes` always includes `#user_id`, `#account_id`, and `#distinct_id`, followed by requested user properties. `#user_id` is an internal association key that customers normally do not care about. In user-facing tables or summaries, show account ID and visitor ID by default and keep `#user_id` only for machine linkage or explicit troubleshooting. If `drilldown_context_id` and the user-event follow-up are present, the Agent may ask for one returned user's event sequence with `analysis drilldown-user-events run|export`.
+- `subject.type=entity`: items contain `entity_value`; do not call user-event commands because non-user entities have no event sequence.
+- When `truncated=true` or complete membership is requested, repeat the same context/source/coordinate with `analysis drilldown-entities export`. Do not page this command.
+- Creating a result cluster is a sibling action from the original analysis cell, not from an entity row. Use the original `query_context_id`, source, and coordinate with `analysis query create-result-cluster` only if that action was advertised.

package/skills/ae-analysis/references/drilldown_events_export.md

@@ -0,0 +1,21 @@
+# analysis drilldown-events export
+
+Stream all raw events behind one selected `EVENT_LIST` coordinate into one `csv.gz` artifact.
+
+Use this only when the selected synchronous preview metric advertises `drilldown_events`. Copy the same `query_context_id`, optional `source`, and semantic `coordinate` used by `analysis drilldown-events run`; never pass raw QP or `target_id`.
+
+```bash
+ae-cli analysis drilldown-events export \
+  --query-context-id <sync_preview_query_context_id> \
+  [--source '{"report_id":1001}'] \
+  --coordinate '<same returned coordinate>' \
+  [--properties '[{"columnName":"<event_property_name>","tableType":"event"}]'] \
+  [--artifact-format csv] \
+  [--timeout-seconds 21600]
+```
+
+This command does not accept `--limit`, `--offset`, `--page-num`, or `--page-size`. Common executes one full-download query and streams its result directly into the artifact; it does not repeatedly call the synchronous preview. “All” follows the platform full-download ceiling (`model_full_download_limit`), not an unbounded query.
+
+When `--properties` is present, the artifact contains required system event columns plus exactly the requested event properties. The export and synchronous preview use the same projection contract.
+
+Inspect `run_id`, wait for completion, then download the bound artifact. The downloaded rows are durable event detail only; they do not create or enlarge the interactive drilldown coordinate set.

package/skills/ae-analysis/references/drilldown_events_run.md

@@ -0,0 +1,23 @@
+# analysis drilldown-events run
+
+Preview raw event rows behind one event-analysis metric cell.
+
+Read [`analysis_drilldown_contract.md`](analysis_drilldown_contract.md) first. Call this command only when the selected event `metric_option` has `analysis_angle=EVENT_LIST` and includes `drilldown_events` in `actions`. Event totals commonly have this angle; triggered-user/entity counts do not.
+
+## Command
+
+```bash
+ae-cli analysis drilldown-events run \
+  --query-context-id <sync_preview_query_context_id> \
+  [--source '{"report_id":1001}'] \
+  --coordinate '{"group_values":["Beijing"],"date":"2026-07-16","metric_index":0}' \
+  [--properties '[{...}]'] \
+  [--limit 100] \
+  [--timeout-seconds 120]
+```
+
+`--query-context-id` must come from the same synchronous preview that returned the selected source and options. `--source` is required only to disambiguate multiple returned sources. Assemble `--coordinate` only by merging the selected returned option fragments. Do not pass `target_id`, raw QP, display-only dates, or values from an export/download.
+
+`--properties` is an optional exact event-property projection. Omit it for default columns; string arrays such as `["#event_time"]` are invalid. Each item uses the backend field names `columnName` and `tableType`, for example `[{"columnName":"<event_property_name>","tableType":"event"}]`. `tableType` uses the documented name `event`, not a numeric enum code. Required system event columns remain present, but unrelated event properties must not be returned.
+
+The response contains at most `limit` event rows from that selected preview cell. `total` is the exact number of matching events, `returned_rows` is the number included in this response, and `truncated=true` means more matching events exist. This command has no paging offset. When complete event detail is required, call `analysis drilldown-events export` once with the same returned context/source/coordinate; that sibling capability uses the backend full-download stream and does not page or enlarge the synchronous selection boundary.

package/skills/ae-analysis/references/drilldown_user_events_export.md

@@ -0,0 +1,23 @@
+# analysis drilldown-user-events export
+
+Export the complete event sequence for one user from a drilldown context. Do not pass raw QP.
+
+## Command
+
+```bash
+ae-cli analysis drilldown-user-events export \
+  --drilldown-context-id <drilldown_context_id> \
+  --user-id <user_id> \
+  [--properties '[{"columnName":"<event_property_name>","tableType":"event"}]'] \
+  [--sort-order desc] \
+  [--artifact-format csv] \
+  [--timeout-seconds 21600]
+```
+
+The nested backend keys are exactly `columnName` and `tableType`, and the table type is the named value `event`. String-name arrays, numeric enum codes, and snake_case rewrites are invalid. With an explicit projection, the artifact retains `#user_id`, account ID, visitor ID, event name, and event time and appends exactly the requested event properties. `#user_id` is internal; Agents should normally present account ID and visitor ID to customers.
+
+Use the exact `drilldown_context_id` and canonical `user_id` returned by a user-subject `analysis drilldown-entities run`. Do not use a custom entity, an entity export artifact, or a guessed identity. Export does not accept `--limit`, `--offset`, `--page-num`, or `--page-size`; Common builds the same authorized event-sequence query without the synchronous 1000-row preview cap and streams one `csv.gz` artifact. The platform full-download ceiling (`model_full_download_limit`) still applies.
+
+Inspect the returned `run_id` with `analysis run inspect`, then download the completed artifact with `analysis artifact download`.
+
+Output is an async run/artifact descriptor. The complete event rows exist only in the downloaded artifact; they cannot be used as new analysis coordinates.

package/skills/ae-analysis/references/drilldown_user_events_run.md

@@ -0,0 +1,34 @@
+# analysis drilldown-user-events run
+
+Query event sequence details for one user after a user-subject `analysis drilldown-entities run`.
+
+Use this only when `analysis drilldown-entities run` returns `subject.type=user`, a `drilldown_context_id`, and the user-event follow-up action. Custom entities do not have event sequences.
+
+Do not pass raw QP.
+
+## Command
+
+```bash
+ae-cli analysis drilldown-user-events run \
+  --drilldown-context-id <drilldown_context_id> \
+  --user-id <user_id> \
+  [--sort-order desc] \
+  [--limit 100] \
+  [--timeout-seconds 120]
+```
+
+## Input
+
+- `--drilldown-context-id`: returned by the same user-subject `analysis drilldown-entities run`.
+- `--user-id`: canonical `user_id` from one item in that same response. Never substitute `entity_value`, `#distinct_id`, account ID, or another identity field.
+- `--properties`: optional exact event-property projection. Omit it to use the default event columns. If present, each item uses backend keys `columnName` and `tableType`, for example `[{"columnName":"<event_property_name>","tableType":"event"}]`. `#user_id`, account ID, visitor ID, event name, and event time remain present; unrelated event properties must not be returned. `#user_id` is an internal association key, so Agents should normally display account ID and visitor ID to customers. Do not pass numeric table-type codes, string-name arrays, or snake_case nested keys.
+- `--sort-order`: `asc` or `desc`.
+- `--event-name-filter`, `--time-filter`, `--time-filter-before-nums`, `--time-filter-after-nums`: optional event sequence filters.
+
+Do not use raw QP, `query_context_id`, or guessed user IDs for this command.
+
+Do not call this command merely because an entity row looks like a user. The explicit subject and follow-up context are the authority.
+
+## Output
+
+The response contains at most `limit` event rows. When `truncated=true`, use `analysis drilldown-user-events export`; do not attempt to fetch another page.