fix(cli): audit tail -n 0 shows nothing instead of the whole ledger

[sturlese]

Bug

flightdeck audit tail -n 0 printed the entire ledger (thousands of entries) instead of zero. On the demo org that's 6731 lines of output for a command that asked for none. A negative -n was wrong too.

Root cause

entries = Ledger(...).entries()[-n:] — the classic negative-slice trap. For n == 0, [-0:] is [0:], i.e. the whole list (-0 == 0). For n == -5, [5:] drops the first five and shows the rest.

Fix

Guard the slice: all_entries[-n:] if n > 0 else [], so asking for 0 (or fewer) entries shows none. Also key the "ledger is empty" notice off the real ledger (all_entries) rather than the truncated view, so tail -n 0 on a populated ledger prints nothing rather than a misleading "ledger is empty".

Tests

• test_audit_tail_n_zero_shows_nothing: after a run, tail -n 0 prints no entries and no false "empty" notice, while tail -n 5 still shows the recent ones.
• test_audit_tail_reports_a_truly_empty_ledger: a genuinely empty ledger still reports "ledger is empty".

The -n 0 test fails on main (prints run_completed) and passes with the fix.

Full suite: 167 passed, 92.65% coverage (gate 85%); ruff clean; offline flightdeck demo + audit verify intact (and audit tail -n 0 now prints 0 lines, was 6731).

Found and fixed by an autonomous bughunt iteration.

[sturlese]

bughunt gate — /code-review high ✅ clean

Reviewed across 8 angles — no correctness or quality findings. The only behavior change is n<=0 on a populated ledger (now prints nothing, correctly); n>0 and the empty-ledger path are unchanged, and the "ledger is empty" notice now keys off the real ledger. Regression test fails on main and passes with the fix. Full local gate: 167 passed, 92.65% coverage, ruff clean, demo intact. Full mode: watching CI to merge on green.