# Setup health diagnostic outside the 244-scenario campaign — report correction r2

This is a presentation correction of [the original health report](https://d1p6rlowie26tp.cloudfront.net/runner-e2e/campaigns/gha-40544678316-1/), not a second eval, hello, attempt or scenario measurement. The retained original result, source139d, Actions run40544678316, authenticated artifact12754606001 and accounting attempt1274c3838148ffc77d79670fc56baaf3 are unchanged. The correction makes the report header use the actual retained model1017 rather than an obsolete publisher profile default.

[Actual Actions run](https://github.com/paperclipai/paperclip/actions/runs/40544678316). One failed setup diagnostic; cleanup passed. This is NOT a canonical scenario measurement, an initial or a correction. It does not change the 244-scenario denominator.

One ordinary automatic lifecycle hello failed: opencode_hello_probe_failed, category UNKNOWN. The working directory, headless permissions, command resolution and validation-via-probe checks were informational; they do not prove the hello succeeded or identify its root cause. No canonical chat was sent:0 issues,0 conversations,0 comments and0 heartbeat runs. Automatic hello provider calls/tokens/cost remain UNKNOWN; zero heartbeat usage is not zero model usage.

Source139d83b7779ea0ec7edb49446221f50719d67d17. Native OpenCode/OpenRouter model openrouter/deepseek/deepseek-v4.1-flash-1017. Actual AWS On-Demand/x86_64 host and pinned PostgreSQL18.6amd64 Docker qualified. Original authenticated artifact/result bytes remain unchanged.

## Working, failed and future coverage

- [Historical partial Codex report](https://d1p6rlowie26tp.cloudfront.net/runner-e2e/campaigns/historical-local-codex-chat-catalog-20261027-158measured-86unrun/):17 original initial passes,141 initial failures across158 attempted IDs;159 original retained measurements. These are historical one-time outcomes, not current OpenCode confidence.
- [Historical pass names, six incomplete journeys and all86 unrun IDs](https://d1p6rlowie26tp.cloudfront.net/runner-e2e/campaigns/historical-local-codex-chat-catalog-20261027-158measured-86unrun/summary.md). Six are **incomplete or blocked journeys**, not complete E2E coverage: access.invite-teammate, access.manage-membership, activity.routine-audit, agents.restore-instructions, budgets-costs.subscription-fees, skills.enable-for-agent. Registration and executable prerequisite probes do not make their intended journeys complete.
- [Original OpenCode backlog initial failure](https://d1p6rlowie26tp.cloudfront.net/runner-e2e/campaigns/gha-40530790741-1/):one actual scenario initial failed setup before canonical chat;243 initials remain unrun. All244 remain behavior-unverified. The outside244 health diagnostic adds no scenario attempt.

The original historical publication is immutable and unchanged; this summary clarifies its earlier registration wording. No old pass is carried into the new model rate. No production/core PR was merged for these measurements.

## Campaign feedback review

0 distinct feedback items await review. Collection gaps and failed submissions remain separate findings.