Data-driven runs: apic run --data rows.csv #53

Why

Newman's --iteration-data and Bruno's CSV runs feed one request many rows. Scenario Outlines already do this for apic test, but a quick "hit this endpoint with these 50 ids" from the CLI has no answer besides a shell loop.

What

• apic run <targets> --data rows.csv|rows.json: one run per row; CSV header names (or JSON object keys) become variables at --var precedence for that iteration.
• Output: text mode prints an iteration 3/50 heading per run; --json adds "iteration": {"index": 3, "total": 50, "row": {...}} to each object.
• --keep-going semantics extend across iterations; the exit code is 1 if any iteration failed. Captures from one iteration do not leak to the next unless --data-share-session.
• --data - reads stdin.

Where

internal/cli/run.go, internal/runner/runner.go (RunAll with per-iteration var overlays), internal/output/render.go.

Docs

docs/cli.md (run flags, JSON field), docs/cheatsheet.md, docs/cookbook.md ("Run one request for every row").

Tests

CSV and JSON inputs; variable precedence; isolation between iterations; exit codes; --json field.