feat: loop metrics report (superseded-PR rate, first-pass green, interventions) #2583

Goal

Measure whether the loop is actually improving delivery rather than assuming it.

Background

Baseline from 2026-10-14: 7 of the last 40 PRs (17.5%) were closed unmerged and recreated. Without a measured baseline and ongoing report, loop quality claims are anecdotal.

Next-release planning review of 2026-10-14/15 against main at 7faf6dfb. Tracking: #2555.

Prerequisites: #2577.

Scope

• Derive from the ledger and GitHub: superseded-PR rate, first-pass CI green rate, human interventions per PR, time from claim to draft PR.
• Text and JSON output; include the baseline for comparison.

Acceptance Criteria

• Report reproduces the 17.5% baseline from repository history.
• Metrics are derived, not hand-entered.
• JSON output is schema-versioned.

Validation

BASE_CLI_SOURCE_DIR=../base-cli/lib/python \
PYTHONPATH=../base-cli/lib/python:lib/python:cli/python \
python -m pytest <new metrics tests>
git diff --check

Non-Goals

• No dashboard.
• No external telemetry.

Project Fields

• Status: Backlog
• Priority: P3
• Area: CLI
• Initiative: Agentic Coding Platform
• Size: S
• Milestone: v1.13.0
• Category: enhancement

Agent Assignment

Assignee: codeforester. Blocked on #2577.