feature: run2 builtin with structured response record

[danieljohnmorris]

Summary

• Adds run2 cmd:t args:L t > R RunResult t - a structured process spawn that returns a typed Record{stdout:t; stderr:t; exit:n} instead of the loose M t t map that run returns
• exit is a number (n), not text, so =0 r.exit, <0 r.exit work without string conversion
• Same no-shell-no-glob security model, same concurrent capture + 10 MiB cap as run; Err only on spawn failure
• run unchanged for compatibility

Repro before/after

Before: accessing exit code required mget m "code" then num code then comparing as number. Loose Map meant no dot-access ergonomics.

After:

r=run2!! "git" ["status" "--short"]
r.stdout   -- text
r.exit     -- 0 or non-zero, already a number
?{=0 r.exit : "clean" ; "dirty"}

What's in the diff

1. add Run2 builtin variant, tree-bridge, and on-wire tag - enum, from_name, name, ALL slice (appended, tag stable), tree-bridge eligible + tree_bridge_returns_result
2. implement run2 interpreter: run_spawn_structured returns typed Record - run_spawn_structured non-wasm + wasm stub, interpreter dispatch, argv validation
3. add run2 verifier: signature table entry and type-check case - BUILTIN_SIGS, type inference match arm, ILO-T013 hints
4. add examples/run-structured.ilo with run:/out: annotations - echo stdout, false exit, true exit, numeric pattern match, nonexistent Err
5. docs: sync run2 to SPEC.md, ai.txt, skill, and CHANGELOG - SPEC Process spawn section updated, skill subsection added, ai.txt regenerated

Test plan

• 6 new unit tests covering: RunResult record shape, stdout text, false exit (1.0), true exit (0.0), stderr capture, exit as Number not Text, nonexistent command returns Err
• examples/run-structured.ilo runs clean in examples_engines harness
• Full suite: 3265 passed, 7 failed (pre-existing AOT worktree-isolation failures, not related to this change)
• cargo fmt --check passes

Follow-ups

• Signal-killed exit is -1.0; could expose the signal number in a future run3 with 4-field record
• stdin pipe arg (run2 cmd args stdin:t) deferred to a follow-up per spec comment in run

[codecov]

❌ 1 Tests Failed:

Tests completed	Failed	Passed	Skipped
4928	1	4927	0

View the full list of 1 ❄️ flaky test(s)

ilo::regression_calendar_arithmetic::add_mo_result_out_of_calendar_range

Flake rate in main: 100.00% (Passed 0 times, Failed 1 times)

Stack Traces | 0.01s run time

thread 'main' (44184) panicked at src/interpreter/mod.rs:1891:21:
attempt to add with overflow
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

To view more test analytics, go to the Test Analytics Dashboard
_{📋 Got 3 mins? Take this short survey to help us improve Test Analytics.}