fix: Codex hang fixes — plan visibility, stdout buffering, reasoning effort (v0.12.4.0) (#536)

* fix: unbuffer Python stdout in codex --json streaming

Python fully buffers stdout when piped (not a TTY). The
`codex exec --json | python3 -c "..."` pattern meant zero output
visible until process exit — users saw nothing for 30+ minutes.

Add PYTHONUNBUFFERED=1 env var, python3 -u flag, and flush=True
to all print() calls in all three Python parser blocks (Challenge,
Consult new session, Consult resumed session).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: per-mode reasoning effort defaults, add --xhigh override

xhigh reasoning uses ~23x more tokens and causes 50+ minute hangs
on large context tasks (OpenAI issues #8545, #8402, #6931).

Per-mode defaults for /codex skill:
- Review: high (bounded diff, needs thoroughness)
- Challenge: high (adversarial but bounded by diff)
- Consult: medium (large context, interactive, needs speed)

Also changes all Outside Voice / adversarial codex invocations
across gstack (resolvers, gen-skill-docs) from xhigh to high.
Users can override with --xhigh flag when they want max reasoning.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: explicit plan content embedding for codex sandbox visibility

Codex runs sandboxed to repo root (-C) and cannot access
~/.claude/plans/. The template already instructed content embedding
but wasn't explicit enough — Claude sometimes shortcut to
referencing the file path, causing Codex to waste 10+ tool calls
searching before giving up.

Strengthen the instruction to make embedding unambiguous: "embed
FULL CONTENT, do NOT reference the file path." Also extract
referenced source file paths from the plan so Codex reads them
directly instead of discovering via rg/find.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: add --xhigh reminder to challenge and consult modes

The --xhigh override was only documented in Step 2A (review).
Steps 2B (challenge) and 2C (consult) lacked the reminder,
so the flag would silently do nothing for those modes.
Found by adversarial review.

* chore: bump version and changelog (v0.12.4.0)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

CHANGELOG.md

@@ -1,5 +1,16 @@
 # Changelog
 
+## [0.12.5.0] - 2026-03-26 — Fix Codex Hangs: 30-Minute Waits Are Gone
+
+Three bugs in `/codex` caused 30+ minute hangs with zero output during plan reviews and adversarial checks. All three are fixed.
+
+### Fixed
+
+- **Plan files now visible to Codex sandbox.** Codex runs sandboxed to the repo root and couldn't see plan files at `~/.claude/plans/`. It would waste 10+ tool calls searching before giving up. Now the plan content is embedded directly in the prompt, and referenced source files are listed so Codex reads them immediately.
+- **Streaming output actually streams.** Python's stdout buffering meant zero output visible until the process exited. Added `PYTHONUNBUFFERED=1`, `python3 -u`, and `flush=True` on every print call across all three Codex modes.
+- **Sane reasoning effort defaults.** Replaced hardcoded `xhigh` (23x more tokens, known 50+ min hangs per OpenAI issues #8545, #8402, #6931) with per-mode defaults: `high` for review and challenge, `medium` for consult. Users can override with `--xhigh` flag when they want maximum reasoning.
+- **`--xhigh` override works in all modes.** The override reminder was missing from challenge and consult mode instructions. Found by adversarial review.
+
 ## [0.12.4.0] - 2026-03-26 — Full Commit Coverage in /ship
 
 When you ship a branch with 12 commits spanning performance work, dead code removal, and test infra, the PR should mention all three. It wasn't. The CHANGELOG and PR summary biased toward whatever happened most recently, silently dropping earlier work.