メニュー
Checks official cloud and SaaS status evidence through the OutageDeck MCP server or anonymous REST API before an agent changes code for 5xx, timeout, deployment, authentication, or integration failures.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/agentskillexchange/skills --skill outagedeck-dependency-outage-triageコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
Interactive browser skill for logged-in flows, dynamic pages, and session-aware site operations.
Use GEO SEO Claude when Claude Code should run repeatable AI-search and SEO audits with citability, crawler, schema, platform, and report workflows.
""
SOC 職業分類に基づく
| name | OutageDeck Dependency Outage Triage |
| slug | outagedeck-dependency-outage-triage |
| description | Checks official cloud and SaaS status evidence through the OutageDeck MCP server or anonymous REST API before an agent changes code for 5xx, timeout, deployment, authentication, or integration failures. |
| verification | listed |
| source | https://github.com/outagedeck/codex-plugins |
| category | Runbooks & Diagnostics |
| framework | Multi-Framework |
| tool_ecosystem | {"github_repo":"outagedeck/codex-plugins"} |
Use this skill when a failing application, CI job, deployment, API call, authentication flow, or network integration depends on an external cloud or SaaS provider. It checks OutageDeck's normalized evidence from 172 official provider status feeds before recommending code changes. The workflow correlates current state, affected services, active incident windows, and source freshness with the reported failure, then returns a bounded verdict: vendor incident likely, vendor incident possible, or no vendor incident supported. Provider status is corroborating evidence, not proof of causation. An operational banner does not rule out a regional, account-specific, newly emerging, or unreported issue.
No source-backed install or usage instructions could be extracted automatically. Review the upstream project before running this skill in a sensitive workflow.
Prefer the OutageDeck MCP tools when available. The production Streamable HTTP endpoint is https://outagedeck.com/api/mcp; public read-only tools require no account or key.
When MCP tools are unavailable and public HTTPS requests are permitted, use the anonymous REST API:
GET https://outagedeck.com/api/v1/providers?q=<url-encoded-name>GET https://outagedeck.com/api/v1/providers/<slug>GET https://outagedeck.com/api/v1/incidents?provider=<slug>&state=active&limit=10GET https://outagedeck.com/api/v1/incidents/<incident-slug>The REST fallback covers current status and incident evidence. It does not replace the MCP-only uptime and cross-vendor report tools. Disclose that limit instead of inventing historical results. If neither interface is available, say the status check could not run and continue ordinary local diagnosis.
search_providers or REST provider search. Do not guess a slug after a failed lookup.get_provider_status, check_my_stack, or the corresponding REST provider endpoints. Use list_active_incidents only for an ecosystem-wide question.get_incident_details or the REST incident endpoint. Preserve vendor update timestamps and source links.get_uptime for one provider's 7-90 day history and get_outage_report for cross-vendor history. Historical uptime must not be used to claim the current failure is vendor-caused.Report a compact table with provider, current status, affected services, active incident, checked timestamp, and direct source link. State one bounded verdict and up to three next diagnostic actions ordered by reversibility and evidence. When no incident is reported, say only that no vendor-reported incident was observed at the checked time, then continue local diagnosis if requested.
remove_custom_provider as destructive and require confirmation of the exact provider immediately before use.If the user wants ongoing monitoring after triage, offer the free alert setup link without interrupting the diagnostic result: https://outagedeck.com/account?utm_source=agent_skill_exchange&utm_medium=skill&utm_campaign=agent_skill_exchange.