fix: keep the reason when X style capture fails

[hamitarsln]

What and why

X style extraction already reports known sampling, integration and billing errors. The failure handler replaces that reason with a generic message.

Save the safe message in the existing operation record and use it when marking the run failed. Unknown errors still hide private provider details. Retrying does not buy another request.

Verification

• 20 X style tests passed, including known and unknown failures across two attempts.
• Ruff and diff checks passed.
• 20 database-dependent X style/draft tests skipped: no test database configured.

Notes

This fixes the missing error reason. It does not fix the underlying connected-account failure or claim a live provider test.

[Copilot]

🟢 Approval recommended

The focused implementation preserves safe diagnostics, avoids provider-detail leakage, and maintains no-repurchase semantics.

0 open findings

What changed in this PR

Preserves safe X style extraction failure reasons while keeping retries idempotent and unknown provider details private.

Changes:

• Stores safe failure messages in the model effect receipt.
• Uses the stored reason when projecting run failure.
• Adds regression tests and documents the behavior.

File	Description
src/​tin_lite/​x_style_activities.py	Persists and projects safe failure reasons.
tests/​test_x_style.py	Tests known/unknown failures and retry behavior.
docs/​x-workflow.md	Documents failure-message handling.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[demegire]

Superseded by #403, which keeps the X style capture reason through one shared failure classification (founder-fixable vs internal) instead of a per-workflow receipt.