Skip to content

RUN_REPORT.md — Portable execution report

A RUN_REPORT captures the outcome of a single execution: of a workflow, a battle, or an evaluation. It is the artifact that lenserfight workflow run, lenserfight battle run, and lenserfight evaluate emit into the user runtime workspace (not .lenserfight/).

Filename

  • Canonical: RUN_REPORT.md
  • Written to: <lenserfight runtime>/workspaces/<workspace_hash>/reports/<slug>-<timestamp>.md
  • A sibling JSON snapshot is written to …/runs/<slug>-<timestamp>.json.

Required frontmatter

KeyTypeNotes
kindrun_reportDiscriminator
schema_versionnumber1
idstringStable id
namestringDisplay name

Common keys: status (ready, blocked, failed), source_ref, started_at, finished_at, cost, latency_ms.

Required sections

  • # Summary
  • # Inputs
  • # Results

Generation

buildWorkflowSimulationReport in automation-objects.ts generates the canonical report skeleton. It is also written by lenserfight workflow run.

Canonical template

yaml
---
kind: run_report
schema_version: 1
id: run_report_<uuid>
slug: workflow-run-report
name: Workflow Run Report
owner: { workspace_id: ws_<uuid> }
visibility: workspace
status: active
version: 0.1.0
---

# Summary
Top-line outcome, cost, and latency.

# Inputs
Describe runtime inputs and references.

# Results
Summarize outputs, failures, and next actions.