Capsule run report

Standard synthesis, conclusion, evidence status and provenance PDF for every completed capsule run.

Version 2 · Status: current

Use when

• A cloud study completes
• A requester pulls or reviews a completed run
• A study is continued or rerun

Requirements

1. Publish a human-readable PDF for each completed exact run, linked from telemetry. The first page presents purpose, synthesis, conclusion and independent evidence status; an attributed appendix preserves the team account and provenance.

2. Keep cloud completion, artifact pulling, independent verification, integration and acceptance distinct. Every claim retains its source and qualification; unknown metrics remain pending.

3. Include stable source session and turn identities, elapsed time when known, generation date, and clearly conditional token-cost estimates distinct from billing.

4. Generate from one reusable template with readable typography, measured wrapping, section hierarchy, non-orphaned headings, page footers and automatic page breaks. PDF generation itself starts no paid model run.

5. Preserve source artifacts and source history. Rerun and continuation requests preserve parent identities and context; no result overwrites an earlier run.

6. Validate generated PDF structure and text extraction. Structural checks do not establish visual acceptance; models do not inspect rendered images.

Implementation sources

• tools/platform-telemetry/report_design.py
• tools/platform-telemetry/delivery.py
• tools/platform-telemetry/syntheses.json