菜单
Query the {{ProductName}} observability stack (OTLP → Grafana/Loki/Tempo/Prometheus) for logs, traces, and metrics during diagnosis. Use when investigating an issue via telemetry — errors, latency, request traces, custom metrics. Preloaded by the rca-investigator agent.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Add a domain entity, aggregate, or value object to the .NET API — rich mutable class with private ctor + static factory (UUIDv7), invariants that throw, tenant_id, identity equality; records for value objects. Use when modeling new domain state under apps/api/{{ProjectName}}.Domain.
Add a Minimal API endpoint (a feature/use case) to the .NET API — feature folder, IEndpoint self-registration, contract validation, Kommand dispatch, Result→ProblemDetails. Use when adding or changing an HTTP endpoint under apps/api/{{ProjectName}}.Api.
Author a CQRS command or query in the {{ProductName}} backend using the Kommand library, the three-scope validation model, and the Result<T> failure pattern. Use whenever creating or modifying a command/query/handler/validator in apps/api/{{ProjectName}}.Application, or wiring its endpoint in {{ProjectName}}.Api.
Choose and wire AI features in the .NET backend — LLM calls, agentic tool-calling/multi-step workflows, RAG/vector search, structured output, streaming. Use when adding or changing anything that talks to a model or orchestrates agents (your agentic pipeline). Tailored to our LLM provider + MAF + pgvector stack.
Scan .NET/C# code for performance anti-patterns (async, memory/strings, collections/LINQ, regex, serialization, I/O) with tiered severity, reporting findings without editing. Use when auditing hot paths, reviewing allocation-heavy code, or analyzing the API for optimization opportunities.
Add or change EF Core (Npgsql/PostgreSQL) persistence in the .NET backend — entities, EF configurations, repositories, queries, migrations. Use when adding a table/entity, a query, a repository, or a migration under apps/api.
| name | query-telemetry |
| description | Query the {{ProductName}} observability stack (OTLP → Grafana/Loki/Tempo/Prometheus) for logs, traces, and metrics during diagnosis. Use when investigating an issue via telemetry — errors, latency, request traces, custom metrics. Preloaded by the rca-investigator agent. |
| argument-hint | <environment> <query> |
| allowed-tools | Read, Bash |
Query our self-hosted observability stack to see what actually happened. Queries the OTLP/Grafana stack:
Loki (logs, LogQL), Tempo (traces, TraceQL), Prometheus (metrics, PromQL). Telemetry is emitted per the
otel-instrumentation skill (sources/meters are {{ProjectName}}.<Feature>).
⚠ Endpoints & auth are PLACEHOLDERS. No observability stack is deployed yet. Fill these in once it exists (your OTLP/Grafana endpoint); until then this is the agreed shape — ask Dan for specifics. Tokens come from env / a vault; never hardcode.
| Environment | Grafana / data-source endpoints | Auth |
|---|---|---|
| prod | Loki <LOKI_URL_PROD> · Tempo <TEMPO_URL_PROD> · Prometheus <PROM_URL_PROD> | <token via env/vault> |
| test | <…_TEST> | <token via env/vault> |
# Logs (LogQL) — e.g. via logcli or the Loki HTTP API:
logcli --addr="$LOKI_URL_PROD" query '{service="{{ProjectName}}-api"} |= "error"' --since=1h -o jsonl
# Metrics (PromQL) — Prometheus HTTP API:
curl -s --get "$PROM_URL_PROD/api/v1/query" --data-urlencode 'query=<PROMQL>'
# Traces (TraceQL) — Tempo HTTP API:
curl -s --get "$TEMPO_URL_PROD/api/search" --data-urlencode 'q=<TRACEQL>'
| Signal | Store / language | Notes |
|---|---|---|
| Logs | Loki / LogQL | Structured logs; correlate by TraceId/SpanId (OTel log correlation). |
| Traces / spans | Tempo / TraceQL | Spans from {{ProjectName}}.<Feature> ActivitySources; HTTP + EF Core auto-instrumentation. |
| Metrics | Prometheus / PromQL | Counters/histograms from {{ProjectName}}.<Feature> meters. |
$0, default prod) and signal (logs/traces/metrics).