Monthly report of precision per reviewer and repository, and cost per fixed finding #48

Background. A scheduled workflow should summarize how useful the bot was each month:

• precision per reviewer and per repository: fixed ÷ (fixed + dismissed), with acknowledged and ignored shown separately
• findings per PR
• total cost split by provider (the Claude API credits vs Bedrock, which is billed to the maintainer's employer)
• cost per fixed finding

Anthropic's usage-report API doesn't cover Bedrock, so spend must come from the engine's own per-call records. For scale: about $1.25 per review on average, roughly $425/month at 340 PRs, of which about $225 lands on Bedrock past the $200 monthly credits.

Current state. No scheduled workflow (.github/workflows has ci, code-review, release and wif-smoke-test). Per-review cost appears only in each summary comment's text.

What to do.

1. Add .github/workflows/monthly-report.yml: schedule: cron '0 6 1 * *' plus workflow_dispatch, with permissions: issues: write, contents: read and GITHUB_TOKEN only (all repos are public, so reading comments needs no other token).
2. tools/report.ts: for each repo where the app is installed, list PRs closed in the previous month, read the outcomes marker (Record each finding's outcome when a pull request closes #47) and the summary comments' cost, and aggregate.
3. Open one issue, "Review report YYYY-MM", with tables per reviewer and per repo, provider split, cost per fixed finding, and the month-over-month change.

Acceptance. A workflow_dispatch run for the current month opens an issue whose totals match a hand count on two PRs.

Related. Depends on #47.

Filed from the 2026-10-31 audit of the NCodeGroup review-bot and project-template design against what was built. This issue stands alone: everything needed to do it is above.