Review follow-ups: P&L doc clarifications and slices.Sort

[tanem]

Small follow-ups from a full-codebase review. Three focused commits; no behaviour change to any command.

Changes

docs: clarify P&L semantics in the README — make three existing behaviours explicit rather than discoverable:

• The exporter owns which deals count (closing deals only; cash flows and entry legs excluded upstream) and the sign convention; the CLI does not re-check them. This is the largest correctness surface, so it's now stated loudly with a "keep the two in step" warning.
• Dates, --last and "today" are interpreted in UTC and deals bucket by UTC day — which can differ from local day from a far-offset timezone (e.g. NZ) near midnight.
• --last runs from N units ago through today inclusive, so 30d covers 31 calendar days.

refactor: sort with slices over the sort package — slices.Sort works on any ordered type, so the int64 less-closure in sort.Slice disappears and sort.Strings becomes slices.Sort. Behaviour unchanged; covered by existing aggregate tests.

docs: explain the decoded-but-unused fields — a comment on OpenPositions/CashFlows records that they're kept for schema fidelity, that cash flows are deliberately excluded from trading P&L, and that open positions are reserved for a future command — so a future tidy-up doesn't delete them and reintroduce a cash-flow bug.

Decisions explicitly not taken

The review also considered a runtime guard on the exporter contract, a --tz flag, changing --last Nd to exactly N days, switching %v→%w, reconciling rounded rows to the total, and extracting a shared-flags helper. Each was deliberately declined as premature or incorrect for this tool; see commit messages / review notes.

Verification

go build, go vet, gofmt -d, and go test ./... all clean after every change.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.65%. Comparing base (6f3129a) to head (2e8661d).

Additional details and impacted files

@@           Coverage Diff           @@
##             main       #3   +/-   ##
=======================================
  Coverage   78.65%   78.65%           
=======================================
  Files          11       11           
  Lines         431      431           
=======================================
  Hits          339      339           
  Misses         75       75           
  Partials       17       17           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.