メニュー
View traces, spans, metrics, feedback, telemetry health, and agent insight reports. Use when the user wants to see traces, check metrics, view top items, submit ratings, diagnose telemetry, or discuss how an agent is doing.
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/BlazeUp-AI/Observal --skill observal-ops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。
Core Observal CLI operations: pull agents into your harness, scan installed components, diagnose and patch harness configs, authenticate, manage CLI settings, and discuss agent insights. Use when the user wants to install an agent, check setup, login, configure the CLI, or ask how an agent is doing.
Advanced Observal operations including session reconciliation, CLI upgrades and downgrades, complete uninstallation, and local fallback mode for offline use. Use when the user wants to reconcile sessions, upgrade or downgrade the CLI, uninstall Observal, or write agent configs locally when the server is unreachable.
Create, update, version, and manage Observal agents. Use when the user wants to create a new agent, update an existing one, release a new version, scaffold a YAML project, add components, build, publish, bulk-create, archive, delete, or restore agents.
SOC 職業分類に基づく
SKILL.md を表示中
| name | observal-ops |
| command | observal |
| description | View traces, spans, metrics, feedback, telemetry health, and agent insight reports. Use when the user wants to see traces, check metrics, view top items, submit ratings, diagnose telemetry, or discuss how an agent is doing. |
| version | 2.1.0 |
| owner | observal |
--output json on every command for stable, machine-readable output.<command> --help first.observal ops overview --output json
observal ops metrics ITEM_NAME --type agent --output json
observal ops metrics ITEM_NAME --type mcp --watch
observal ops top --type agent --output json
observal ops top --type mcp --output json
observal ops traces --limit 20 --output json
observal ops traces --platform kiro --days 7 --output json
observal ops traces --turn --limit 5
observal ops traces --span --limit 3
observal ops spans TRACE_ID --output json
observal ops feedback ITEM_NAME --type mcp --output json
observal ops rate MCP_NAME --stars 5 --type mcp --comment 'Worked great'
observal ops rate AGENT_NAME --stars 4 --type agent
--stars (1-5) and --type are required. --comment is optional.
observal ops telemetry status
observal ops telemetry test
status is the reliable check: it queries server event counts and local SQLite buffer. test may return 404 on newer servers (legacy endpoint). If status shows events flowing, telemetry is healthy.
Diagnosis: status OK → healthy. No events → check observal auth status. Server reachable but no events → hooks not installed, suggest observal doctor.
Use this when the user asks how an agent is doing, what changed, why a version regressed, what to improve, or wants to talk through an insight report.
Start with machine readable reports:
observal ops insights list AGENT_NAME --output json
observal ops insights show AGENT_NAME latest --output json
Fetch one section when the user asks a narrow question:
observal ops insights show AGENT_NAME latest --section at_a_glance --output json
observal ops insights show AGENT_NAME latest --section what_they_work_on --output json
observal ops insights show AGENT_NAME latest --section interaction_style --output json
observal ops insights show AGENT_NAME latest --section usage_patterns --output json
observal ops insights show AGENT_NAME latest --section what_works --output json
observal ops insights show AGENT_NAME latest --section friction_analysis --output json
observal ops insights show AGENT_NAME latest --section suggestions --output json
observal ops insights show AGENT_NAME latest --section usage_cost_analysis --output json
observal ops insights show AGENT_NAME latest --section version_comparison --output json
observal ops insights show AGENT_NAME latest --section regression_detection --output json
observal ops insights show AGENT_NAME latest --section on_the_horizon --output json
observal ops insights show AGENT_NAME latest --section fun_ending --output json
Section meanings:
| Section | Use for |
|---|---|
at_a_glance | Overall health, working areas, blockers, quick win |
what_they_work_on | Project areas and session counts |
interaction_style | User behavior and collaboration pattern |
usage_patterns | Session length, tool distribution, prompts |
what_works | Agent strengths and evidence |
friction_analysis | Recurring failures, severity, examples |
suggestions | Config changes, features, prompts, habits |
usage_cost_analysis | Cost, cache, model efficiency |
version_comparison | Current version versus baseline |
regression_detection | Improvements or degradations over time |
on_the_horizon | Higher leverage next workflows |
fun_ending | Memorable qualitative moment |
If no completed report exists, generate one:
observal ops insights generate AGENT_NAME --period 14 --wait
For versioned analysis, request or infer versions from list, then generate or show version scoped reports:
observal ops insights generate AGENT_NAME --version 1.2.0 --compare 1.1.0 --period 30 --wait
observal ops insights show AGENT_NAME latest --output json
Answer like an analyst: cite the report period, session count, strengths, friction, cost, version notes, and two or three concrete next actions. If data is thin, say so.