Support Console Ops
Resolve telecom operations tasks by querying the support console REST API and
mapping evidence to structured JSON decisions. The API is a read-only
exploration surface; the skill covers five task types.
Quick Start
- Read the prompt to identify the task type and the payload file(s).
- Parse the payload(s) to build the work list (tickets, cases, or incidents).
- Query the API iteratively, gathering evidence for each item.
- Map evidence to decisions following the rules in this skill.
- Return JSON that conforms to the provided
answer_template.json.
All API endpoints return JSON arrays of objects or single objects. Use curl
with -s and pipe through python3 -m json.tool for readability when exploring,
but prefer direct JSON parsing for decision logic.
API Reference
See references/api_reference.md for full
endpoint descriptions, field shapes, and interpretation rules. Load it when
the task involves unfamiliar endpoints.
Key base endpoints by task type:
| Task Type |
Primary Endpoints |
| Ticket batch |
/api/tickets, /api/accounts, /api/outages, /api/diagnostics/<ticket_id>, /api/troubleshooting/<ticket_id> |
| Contact center cases |
/api/cases, /api/customers, /api/lines, /api/devices/<device_id>, /api/plans/<plan_id>, /api/bills |
| Enterprise incident response |
/api/enterprise/incidents, /api/enterprise/accounts, /api/enterprise/export-runs, /api/enterprise/messages, /api/enterprise/sla/<account_id> |
| Mobile data recovery |
/api/cases, /api/lines, /api/devices/<device_id>, /api/plans/<plan_id>, /api/bills |
Task Type Decision Tree
Read the prompt to determine the task:
- "support operations analyst" + "batch" + CSV payload → Ticket Batch Resolution
- "contact-center lead" + case queue JSON → Contact Center Case Routing
- "enterprise support lead" + complaint email → Enterprise Incident Response
- "queue-quality analyst" + queue snapshot CSV → Ticket Queue Classification
- "mobile-data recovery analyst" + worklist JSON → Mobile Data Recovery
Ticket Batch Resolution
Read each ticket from the CSV payload. For each ticket:
- Fetch the ticket via
GET /api/tickets/<ticket_id> — get account_id, service_area, service_type.
- Fetch the account via
GET /api/accounts/<account_id>:
- If HTTP 404 or
account_id starts with a non-standard prefix → FAILED / INVALID_ACCOUNT.
- If
status: "Suspended" → FAILED / INELIGIBLE_ACCOUNT.
- If
authentication.last_login_status is not "SUCCESS" → FAILED / AUTH_FAILED.
- Fetch outages via
GET /api/outages and match: active: true, same service_area, service_types includes ticket’s service_type. If match exists:
- Status →
PENDING_ACTION, route → OUTAGE_WAIT, outage_id → matched outage’s ID.
diagnostic_needed → false, all issue flags → false.
- No active outage → fetch diagnostics via
GET /api/diagnostics/<ticket_id>.
- Examine
root_causes. Ignore GENERATED_NOISE entries.
CONFIGURATION_DRIFT → fetch troubleshooting via GET /api/troubleshooting/<ticket_id>.
- If troubleshooting has
steps and post_bandwidth_mbps near subscribed_mbps → RESOLVED / AUTO_TROUBLESHOOTING.
FIBER_DROP_DAMAGE or SIGNAL_LOSS → ESCALATED / ESCALATION to FIELD_OPS.
NETWORK_CAPACITY → ESCALATED to NETWORK_ENGINEERING.
PROVISIONING_STALE → ESCALATED to TIER2_SUPPORT.
- Issue flags: set
latency_issue, stability_issue, bandwidth_issue based on diagnostics:
- Compare
bandwidth_mbps to subscribed_mbps (substantially lower → bandwidth_issue: true).
- Compare
latency_ms (>80 ms → latency_issue: true).
- Compare
jitter_ms (>25 ms → stability_issue: true).
diagnostic_needed is true when diagnostics were fetched and used.
Batch summary: count each final_resolution_status category plus tickets with
PENDING_ACTION as tickets_requiring_customer_wait.
Ticket Queue Classification
Read tickets from a CSV with ticket_id, account_id, reported_service_type,
queue_note. Classify each:
- Active outage? →
PENDING_ACTION, key_blocker: ACTIVE_OUTAGE, no escalation needed.
- Invalid account? →
FAILED, key_blocker: INVALID_ACCOUNT.
- Auth failure? →
FAILED, key_blocker: AUTH_FAILED.
- Overdue suspension? →
FAILED, key_blocker: OVERDUE_SUSPENSION, route to ACCOUNTS_PAYABLE.
- Fraud suspension? →
FAILED, key_blocker: FRAUD_SUSPENSION.
- Config/provisioning drift? →
RESOLVED when troubleshooting succeeds; diagnostic_required: true.
- Network capacity / backbone? →
ESCALATED to NETWORK_ENGINEERING, key_blocker: NETWORK_CAPACITY.
- Physical line fault? →
ESCALATED to FIELD_OPS, key_blocker: PHYSICAL_LINE_FAULT.
- Provisioning stale? →
ESCALATED to TIER2_SUPPORT, key_blocker: PROVISIONING_STALE.
Where diagnostics or troubleshooting confirms a resolution, set
diagnostic_required: true. Where the blocker is self-evident (outage, invalid
account, etc.), set diagnostic_required: false.
Contact Center Case Routing
Read each case from case_queue.json. For each case:
- Fetch the case to get
line_id, device_id, customer_id, issue_type, customer_location.
- Fetch the line via
GET /api/lines/<line_id>:
- If
status: "Suspended" with suspension_reason: "OVERDUE_BILL":
- Primary action →
SEND_PAYMENT_REQUEST.
- Secondary action →
RESUME_LINE_REBOOT.
- Fetch
/api/bills, filter by customer_id, find overdue bill.
- Set
bill_id and charge_amount_usd from the bill.
final_route → BILLING_RECOVERY.
- If
roaming_enabled: true but case is abroad → fetch device; if phone_roaming_enabled: false → TOGGLE_ROAMING, SELF_SERVICE. If phone_roaming_enabled: true but still no data → ENABLE_LINE_ROAMING, CARRIER_UPDATE.
- Fetch the device via
GET /api/devices/<device_id>:
- Map device state to action:
sim_status: "missing" + signal_strength: "none" → RESEAT_SIM.
phone_roaming_enabled: false while abroad → TOGGLE_ROAMING.
data_saver_mode: true with slow data → TOGGLE_DATA_SAVER.
vpn_connected: true with slow data → DISCONNECT_VPN.
mobile_data_enabled: false with no data → TOGGLE_MOBILE_DATA.
can_send_mms: false with MMS issue → GRANT_MESSAGING_PERMISSION.
- Outdated
network_mode_preference → SET_NETWORK_MODE.
- Permission field: For
GRANT_MESSAGING_PERMISSION, check messaging_permissions. If storage is true but can_send_mms is false, set permission: "storage". Otherwise permission: "NONE".
- Default:
NO_ACTION for both primary and secondary when no rule matches.
Mobile Data Recovery
Read each case from the worklist JSON. For each case:
- Fetch the case, line, and device as in Contact Center routing.
- Check if data refuel is needed:
- Fetch the plan via
GET /api/plans/<plan_id>.
- If
data_used_gb > plan.data_limit_gb or the case reports "Data stopped after usage limit":
- Primary action →
REFUEL_DATA.
- Use customer preferences for
data_refuel_gb if provided.
charge_amount_usd = data_refuel_gb * plan.data_refueling_price_per_gb.
final_route → DATA_RECOVERY.
carrier_update_required → false.
- Device-based actions:
phone_roaming_enabled: false while abroad → TOGGLE_ROAMING or ENABLE_LINE_ROAMING. If line-side roaming also needs enabling: carrier_update_required: true, final_route: CARRIER_UPDATE.
data_saver_mode: true → TOGGLE_DATA_SAVER, DEVICE_SETTING_FIX.
mobile_data_enabled: false → TOGGLE_MOBILE_DATA, DEVICE_SETTING_FIX.
- Outdated
network_mode_preference → SET_NETWORK_MODE, DEVICE_SETTING_FIX.
- Worklist summary: count
data_refuel_cases, carrier_updates, device_setting_fixes, human_transfers, and sum total_estimated_customer_charge_usd.
Enterprise Incident Response
Given a complaint email and response requirements:
- Identify the incident using the approximate reference from the complaint.
GET /api/enterprise/incidents/<incident_id> to extract enterprise_account_id,
severity, engineering_owner, account_owner.
- Fetch the enterprise account via
GET /api/enterprise/accounts/<account_id>.
- Fetch export runs via
GET /api/enterprise/export-runs. Filter by
enterprise_account_id. Identify consecutive FAILED runs to determine
failure_window (start_date, end_date, failed_days). The failure_code on
those runs becomes root_cause_category. Count SUCCEEDED runs after the
failure window → backfill_days.
- Fetch messages via
GET /api/enterprise/messages. Look for messages by
the engineering owner and account owner. Extract root cause explanation from
message bodies. If a message about the root cause is in a channel whose
name contains "archive" → contributing_alert_issue: ARCHIVED_ALERT_ROUTE.
Otherwise NONE.
- Fetch SLA via
GET /api/enterprise/sla/<enterprise_account_id>.
monthly_export_credit_percent → sla_credit_percent.
- Name derivation from naming_style requirements:
channel_name: lowercase, hyphenated client name.
evidence_folder: "Client Name Month Year Investigation".
report_title: "Client Name Export Failure - Resolution Report".
- Share permissions: assign
view or edit to users per requirements.
- Response status: If
sla_credit_percent > 0 → NEEDS_FINANCE_REVIEW.
Otherwise READY_TO_SEND.
General Rules
- Always preserve payload order when returning JSON arrays.
- Use
"NONE" for enum fields when no applicable value exists.
- Use
"" for string fields when no value applies.
- Use
0.0 for numeric fields when no value applies.
- Map
NO_ACTION as secondary action when only one action is needed.
- When an API call returns an HTTP error, the item is failed with an appropriate
key_blocker or final_resolution_status of FAILED.
- All monetary values use exactly two decimal places. Data refuel GB uses one decimal.
- Read
answer_template.json before starting to understand the exact output schema.
1---2name: support-console-ops-33description: Resolve telecom support tasks using a shared support console REST API. Use when a prompt provides TASK_ENV_BASE_URL and asks to resolve support tickets (batch or queue), handle contact-center cases, classify ticket queues, prepare enterprise export incident responses, or recover mobile data. Covers ticket resolution, case routing, queue classification, enterprise incident response, and mobile data recovery workflows.4metadata:5 source: https://skillmd.com/skills/prism-shadow/support-console-ops-36---78# Support Console Ops910Resolve telecom operations tasks by querying the support console REST API and11mapping evidence to structured JSON decisions. The API is a read-only12exploration surface; the skill covers five task types.1314## Quick Start15161. Read the prompt to identify the task type and the payload file(s).172. Parse the payload(s) to build the work list (tickets, cases, or incidents).183. Query the API iteratively, gathering evidence for each item.194. Map evidence to decisions following the rules in this skill.205. Return JSON that conforms to the provided `answer_template.json`.2122All API endpoints return JSON arrays of objects or single objects. Use `curl`23with `-s` and pipe through `python3 -m json.tool` for readability when exploring,24but prefer direct JSON parsing for decision logic.2526## API Reference2728See [references/api_reference.md](references/api_reference.md) for full29endpoint descriptions, field shapes, and interpretation rules. Load it when30the task involves unfamiliar endpoints.3132Key base endpoints by task type:3334| Task Type | Primary Endpoints |35|---|---|36| Ticket batch | `/api/tickets`, `/api/accounts`, `/api/outages`, `/api/diagnostics/<ticket_id>`, `/api/troubleshooting/<ticket_id>` |37| Contact center cases | `/api/cases`, `/api/customers`, `/api/lines`, `/api/devices/<device_id>`, `/api/plans/<plan_id>`, `/api/bills` |38| Enterprise incident response | `/api/enterprise/incidents`, `/api/enterprise/accounts`, `/api/enterprise/export-runs`, `/api/enterprise/messages`, `/api/enterprise/sla/<account_id>` |39| Mobile data recovery | `/api/cases`, `/api/lines`, `/api/devices/<device_id>`, `/api/plans/<plan_id>`, `/api/bills` |4041## Task Type Decision Tree4243Read the prompt to determine the task:4445- **"support operations analyst" + "batch" + CSV payload** → Ticket Batch Resolution46- **"contact-center lead" + case queue JSON** → Contact Center Case Routing47- **"enterprise support lead" + complaint email** → Enterprise Incident Response48- **"queue-quality analyst" + queue snapshot CSV** → Ticket Queue Classification49- **"mobile-data recovery analyst" + worklist JSON** → Mobile Data Recovery5051## Ticket Batch Resolution5253Read each ticket from the CSV payload. For each ticket:54551. **Fetch the ticket** via `GET /api/tickets/<ticket_id>` — get `account_id`, `service_area`, `service_type`.562. **Fetch the account** via `GET /api/accounts/<account_id>`:57 - If HTTP 404 or `account_id` starts with a non-standard prefix → `FAILED` / `INVALID_ACCOUNT`.58 - If `status: "Suspended"` → `FAILED` / `INELIGIBLE_ACCOUNT`.59 - If `authentication.last_login_status` is not `"SUCCESS"` → `FAILED` / `AUTH_FAILED`.603. **Fetch outages** via `GET /api/outages` and match: `active: true`, same `service_area`, `service_types` includes ticket’s `service_type`. If match exists:61 - Status → `PENDING_ACTION`, route → `OUTAGE_WAIT`, `outage_id` → matched outage’s ID.62 - `diagnostic_needed` → `false`, all issue flags → `false`.634. **No active outage** → fetch diagnostics via `GET /api/diagnostics/<ticket_id>`.64 - Examine `root_causes`. Ignore `GENERATED_NOISE` entries.65 - `CONFIGURATION_DRIFT` → fetch troubleshooting via `GET /api/troubleshooting/<ticket_id>`.66 - If troubleshooting has `steps` and `post_bandwidth_mbps` near `subscribed_mbps` → `RESOLVED` / `AUTO_TROUBLESHOOTING`.67 - `FIBER_DROP_DAMAGE` or `SIGNAL_LOSS` → `ESCALATED` / `ESCALATION` to `FIELD_OPS`.68 - `NETWORK_CAPACITY` → `ESCALATED` to `NETWORK_ENGINEERING`.69 - `PROVISIONING_STALE` → `ESCALATED` to `TIER2_SUPPORT`.705. **Issue flags:** set `latency_issue`, `stability_issue`, `bandwidth_issue` based on diagnostics:71 - Compare `bandwidth_mbps` to `subscribed_mbps` (substantially lower → `bandwidth_issue: true`).72 - Compare `latency_ms` (>80 ms → `latency_issue: true`).73 - Compare `jitter_ms` (>25 ms → `stability_issue: true`).746. **`diagnostic_needed`** is `true` when diagnostics were fetched and used.7576Batch summary: count each `final_resolution_status` category plus tickets with77`PENDING_ACTION` as `tickets_requiring_customer_wait`.7879## Ticket Queue Classification8081Read tickets from a CSV with `ticket_id`, `account_id`, `reported_service_type`,82`queue_note`. Classify each:83841. **Active outage?** → `PENDING_ACTION`, `key_blocker: ACTIVE_OUTAGE`, no escalation needed.852. **Invalid account?** → `FAILED`, `key_blocker: INVALID_ACCOUNT`.863. **Auth failure?** → `FAILED`, `key_blocker: AUTH_FAILED`.874. **Overdue suspension?** → `FAILED`, `key_blocker: OVERDUE_SUSPENSION`, route to `ACCOUNTS_PAYABLE`.885. **Fraud suspension?** → `FAILED`, `key_blocker: FRAUD_SUSPENSION`.896. **Config/provisioning drift?** → `RESOLVED` when troubleshooting succeeds; `diagnostic_required: true`.907. **Network capacity / backbone?** → `ESCALATED` to `NETWORK_ENGINEERING`, `key_blocker: NETWORK_CAPACITY`.918. **Physical line fault?** → `ESCALATED` to `FIELD_OPS`, `key_blocker: PHYSICAL_LINE_FAULT`.929. **Provisioning stale?** → `ESCALATED` to `TIER2_SUPPORT`, `key_blocker: PROVISIONING_STALE`.9394Where diagnostics or troubleshooting confirms a resolution, set95`diagnostic_required: true`. Where the blocker is self-evident (outage, invalid96account, etc.), set `diagnostic_required: false`.9798## Contact Center Case Routing99100Read each case from `case_queue.json`. For each case:1011021. **Fetch the case** to get `line_id`, `device_id`, `customer_id`, `issue_type`, `customer_location`.1032. **Fetch the line** via `GET /api/lines/<line_id>`:104 - If `status: "Suspended"` with `suspension_reason: "OVERDUE_BILL"`:105 - Primary action → `SEND_PAYMENT_REQUEST`.106 - Secondary action → `RESUME_LINE_REBOOT`.107 - Fetch `/api/bills`, filter by `customer_id`, find overdue bill.108 - Set `bill_id` and `charge_amount_usd` from the bill.109 - `final_route` → `BILLING_RECOVERY`.110 - If `roaming_enabled: true` but case is abroad → fetch device; if `phone_roaming_enabled: false` → `TOGGLE_ROAMING`, `SELF_SERVICE`. If `phone_roaming_enabled: true` but still no data → `ENABLE_LINE_ROAMING`, `CARRIER_UPDATE`.1113. **Fetch the device** via `GET /api/devices/<device_id>`:112 - Map device state to action:113 - `sim_status: "missing"` + `signal_strength: "none"` → `RESEAT_SIM`.114 - `phone_roaming_enabled: false` while abroad → `TOGGLE_ROAMING`.115 - `data_saver_mode: true` with slow data → `TOGGLE_DATA_SAVER`.116 - `vpn_connected: true` with slow data → `DISCONNECT_VPN`.117 - `mobile_data_enabled: false` with no data → `TOGGLE_MOBILE_DATA`.118 - `can_send_mms: false` with MMS issue → `GRANT_MESSAGING_PERMISSION`.119 - Outdated `network_mode_preference` → `SET_NETWORK_MODE`.1204. **Permission field:** For `GRANT_MESSAGING_PERMISSION`, check `messaging_permissions`. If `storage` is `true` but `can_send_mms` is `false`, set `permission: "storage"`. Otherwise `permission: "NONE"`.1215. **Default:** `NO_ACTION` for both primary and secondary when no rule matches.122123## Mobile Data Recovery124125Read each case from the worklist JSON. For each case:1261271. **Fetch the case, line, and device** as in Contact Center routing.1282. **Check if data refuel is needed:**129 - Fetch the plan via `GET /api/plans/<plan_id>`.130 - If `data_used_gb > plan.data_limit_gb` or the case reports "Data stopped after usage limit":131 - Primary action → `REFUEL_DATA`.132 - Use customer preferences for `data_refuel_gb` if provided.133 - `charge_amount_usd = data_refuel_gb * plan.data_refueling_price_per_gb`.134 - `final_route` → `DATA_RECOVERY`.135 - `carrier_update_required` → `false`.1363. **Device-based actions:**137 - `phone_roaming_enabled: false` while abroad → `TOGGLE_ROAMING` or `ENABLE_LINE_ROAMING`. If line-side roaming also needs enabling: `carrier_update_required: true`, `final_route: CARRIER_UPDATE`.138 - `data_saver_mode: true` → `TOGGLE_DATA_SAVER`, `DEVICE_SETTING_FIX`.139 - `mobile_data_enabled: false` → `TOGGLE_MOBILE_DATA`, `DEVICE_SETTING_FIX`.140 - Outdated `network_mode_preference` → `SET_NETWORK_MODE`, `DEVICE_SETTING_FIX`.1414. **Worklist summary:** count `data_refuel_cases`, `carrier_updates`, `device_setting_fixes`, `human_transfers`, and sum `total_estimated_customer_charge_usd`.142143## Enterprise Incident Response144145Given a complaint email and response requirements:1461471. **Identify the incident** using the approximate reference from the complaint.148 `GET /api/enterprise/incidents/<incident_id>` to extract `enterprise_account_id`,149 `severity`, `engineering_owner`, `account_owner`.1502. **Fetch the enterprise account** via `GET /api/enterprise/accounts/<account_id>`.1513. **Fetch export runs** via `GET /api/enterprise/export-runs`. Filter by152 `enterprise_account_id`. Identify consecutive `FAILED` runs to determine153 `failure_window` (start_date, end_date, failed_days). The `failure_code` on154 those runs becomes `root_cause_category`. Count `SUCCEEDED` runs after the155 failure window → `backfill_days`.1564. **Fetch messages** via `GET /api/enterprise/messages`. Look for messages by157 the engineering owner and account owner. Extract root cause explanation from158 message bodies. If a message about the root cause is in a channel whose159 name contains "archive" → `contributing_alert_issue: ARCHIVED_ALERT_ROUTE`.160 Otherwise `NONE`.1615. **Fetch SLA** via `GET /api/enterprise/sla/<enterprise_account_id>`.162 `monthly_export_credit_percent` → `sla_credit_percent`.1636. **Name derivation** from naming_style requirements:164 - `channel_name`: lowercase, hyphenated client name.165 - `evidence_folder`: "Client Name Month Year Investigation".166 - `report_title`: "Client Name Export Failure - Resolution Report".1677. **Share permissions:** assign `view` or `edit` to users per requirements.1688. **Response status:** If `sla_credit_percent > 0` → `NEEDS_FINANCE_REVIEW`.169 Otherwise `READY_TO_SEND`.170171## General Rules172173- Always preserve payload order when returning JSON arrays.174- Use `"NONE"` for enum fields when no applicable value exists.175- Use `""` for string fields when no value applies.176- Use `0.0` for numeric fields when no value applies.177- Map `NO_ACTION` as secondary action when only one action is needed.178- When an API call returns an HTTP error, the item is failed with an appropriate179 `key_blocker` or `final_resolution_status` of `FAILED`.180- All monetary values use exactly two decimal places. Data refuel GB uses one decimal.181- Read `answer_template.json` before starting to understand the exact output schema.