Restore exhaustive file-conversion field matching

[patch-goblin]

Review feedback

Addresses the field-matching comment on #1852.

• Restore the explicit AsyncApiCallOutput::FileConversion field pattern instead of .., so SDK field additions require a compiler-checked decision.
• Keep the mutable outputs/status bindings and all formatting/download behavior unchanged.

Validation

• nix develop --command cargo test --locked --test output_formats api_call_status_honors_format_including_downloads: passes, including the existing format/config precedence and completed-download cases.
• nix develop --command cargo fmt -- --check: passes.
• git diff --check: passes.

No new tests are needed for this structural restoration; compilation checks the exhaustive pattern, and the existing regression verifies that runtime behavior is unchanged.

Relationship to #1852

This follow-up targets #1852's branch, so its diff contains only the review correction. The original PR and branch are unchanged: this run's write rules allow new branches and draft PRs, not updates to existing PR branches. The new branch was created from current main, then fast-forwarded through #1852's existing commit before applying this correction.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 66.79%. Comparing base (658178c) to head (f64966f).

Additional details and impacted files

@@                                   Coverage Diff                                    @@
##           patch-goblin/slack/20645017-f05b-46ae-9c57-3ddf0c476d48    #1853   +/-   ##
========================================================================================
  Coverage                                                    66.79%   66.79%           
========================================================================================
  Files                                                           43       43           
  Lines                                                         8456     8456           
========================================================================================
  Hits                                                          5648     5648           
  Misses                                                        2808     2808           

Flag	Coverage Δ
unittests	66.79% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.

[patch-goblin]

Token cost associated with this merged PR: $27.26226 across 1 Codex command. This total includes any recorded usage from its originating Slack run.

Patch Goblin run details

• PR: #1853
• Head branch: patch-goblin/slack/bcf422b3-3286-4a2e-ab6b-fe9efcc2d145
• Base branch: patch-goblin/slack/20645017-f05b-46ae-9c57-3ddf0c476d48
• Runs scanned: 1
• Runs with token usage: 1
• Codex commands: 1
• Input tokens: 10,398,012
• Cache read tokens: 10,227,975
• Cache write tokens: 162,924
• Uncached input tokens: 7,113
• Output tokens: 34,546
• Reasoning output tokens: 13,986
• Total tokens: 10,432,558
• Estimated cost: $27.26226

Model	Calls	Input	Cache reads	Cache writes	Output	Total	Estimated cost
gpt-6-astra	1	10,398,012	10,227,975	162,924	34,546	10,432,558	$27.26226

Pricing uses built-in standard API rates for recognized Codex models. Cache reads and writes are priced separately from uncached input when the model publishes distinct rates, and long-context multipliers are applied for recognized models that publish them.

For Slack-origin pull requests, any recorded usage from the originating Slack run is included in full. If that run produced pull requests in multiple repositories, the shared Slack portion appears in each pull request's total and must not be summed more than once.