Prepare, submit and reconcile reviewed Fusion Automation jobs with exact sources, scoped authorization and explicit budgets. Not for arbitrary model-generated cloud scripts.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Inspect fusion_cloud_status, authorization owner, reviewed recipes, scope and admission budget. Never reuse native MCP tokens, distribute a client secret or request a PAT.
Choose an installed recipe and declared typed inputs. Freeze exact source versions/configuration/hashes. A generic signed activity does not constrain code, data or cost.
Use fusion_cloud_job_prepare to resolve actual dependencies and reservation. Required hard-cap or immutability guarantees remain blockers when unsupported; do not weaken them silently.
For a bounded batch, use fusion_cloud_batch_prepare with one approved recipe, frozen common source context, approved object-storage destination, a stable request key and explicit unique variant IDs. Review the whole prepared mapping before submission. fusion_cloud_batch_resume admits only the requested bounded wave of unattempted children; it does not refresh expiry, bypass per-job budgets or publish into Fusion/ERP. Use fusion_cloud_batch_inspect after interruption, and retain each child job ID for detailed status, output validation and billing.
Submit the unchanged plan with an existing compute grant and unique idempotency key. Durable intent/reservation precedes submission. Lost acknowledgments and unknown states never justify blind resubmission.
Poll the plugin job ID. Provider success remains validating until trusted output checks pass. Cancel only where qualified; cancellation is not rollback or free compute. Final billing comes from the trusted reconciliation adapter; unknown exposure retains its reservation.
Unknown validation intent or a recorded output-identity collision blocks the batch. Do not clear records, regenerate keys, revalidate to erase a conflict or submit remaining variants as a workaround. A receipt ID collision check does not prove physical storage-key isolation; that is part of the trusted stager's qualification.
Report processing, validation, artifact and cost evidence separately. Offline mocks do not establish geometry, residency, metering or service compatibility. See enterprise recipes and batch recovery.
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Analyze reviewed Grafana log evidence for event patterns, error-rate changes, and enrolled security markers. Use when a user needs bounded log analytics for an incident window, error-log comparison, anomaly triage, operational debugging, or a cautious check of approved security signals.
日本語の概要は準備中です。原文の説明を表示しています。
Run a read-only App Store-readiness preflight for a declared iOS target without claiming approval or submitting anything.
日本語の概要は準備中です。原文の説明を表示しています。
Audit whether a declared local review evaluation can support fair quality claims. Use only when the user explicitly asks to audit review-eval validity or use the bundled synthetic fixture.
日本語の概要は準備中です。原文の説明を表示しています。
Deterministically score comparable imported review-run lanes with root-cause metrics and uncertainty. Use only when the user explicitly asks to benchmark normalized local review runs or use the bundled synthetic fixture.
日本語の概要は準備中です。原文の説明を表示しています。
Stage and verify Fusion exports, drawing PDFs, flat-pattern DXF, viewport images and local renders. Use for engineering deliverables, not automatic publication or release.
日本語の概要は準備中です。原文の説明を表示しています。
Prepare, execute and verify authorized Fusion CAD changes using resolved entities, explicit units and state-bound plans. Use for reviewed parameter, sketch, feature, assembly or configuration operations.
日本語の概要は準備中です。原文の説明を表示しています。