Preserve native media and expose transport input support

[senamakel]

Native attachment routing needs separate facts for the selected model and its transport. This adds conservative ChatModel::supports_input checks, preserves discovered input modalities, and serializes images in OpenAI Responses with PDF input enabled only when explicitly confirmed. Unsupported media and malformed inline Base64 fail before transport instead of being silently dropped.

Validation (all pass, run from this repository root):

cargo test --workspace --all-features
cargo clippy --workspace --all-targets --all-features -- -D warnings
cargo fmt --all -- --check

The LLM suite has 568 unit tests, one integration test, and one doctest. New regressions cover tied and unequal context windows, unknown and empty modality facts, root-level capability exports, modality-only discovery supplementation, pure native-media serialization, and public invoke/stream rejection before I/O. New successful-media tests use synthetic payloads; no socket fixture is added.

Dependency for tinyhumansai/openhuman#6964.

[tinysweeper]

Tiny Sweeper review

⚠️ Review failed for 84d4cafa51ee. the review of #61 did not finish within 900s

[coderabbitai]

Warning

Review limit reached

• Run on-demand review

_{This review includes 28 billable files and costs up to $7.00.}

Or wait 5 minutes for your next included review.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 85b53113-f84e-425b-8a2e-c53e7dc08a18

📥 Commits

Reviewing files that changed from the base of the PR and between e7a2599 and 84d4caf.

⛔ Files ignored due to path filters (1)

• Cargo.lock is excluded by !**/*.lock

📒 Files selected for processing (28)

• crates/tinyinference-llm/Cargo.toml
• crates/tinyinference-llm/src/catalog/mod_tests.rs
• crates/tinyinference-llm/src/catalog/parsing.rs
• crates/tinyinference-llm/src/catalog/types.rs
• crates/tinyinference-llm/src/lib.rs
• crates/tinyinference-llm/src/model/decorators.rs
• crates/tinyinference-llm/src/model/discover/cache.rs
• crates/tinyinference-llm/src/model/discover/cache_tests.rs
• crates/tinyinference-llm/src/model/discover/fetch.rs
• crates/tinyinference-llm/src/model/discover/fetch_tests.rs
• crates/tinyinference-llm/src/model/discover/mod.rs
• crates/tinyinference-llm/src/model/discover/parse.rs
• crates/tinyinference-llm/src/model/discover/parse_tests.rs
• crates/tinyinference-llm/src/model/discover/types.rs
• crates/tinyinference-llm/src/model/migration_tests.rs
• crates/tinyinference-llm/src/model/types.rs
• crates/tinyinference-llm/src/providers/anthropic/mod.rs
• crates/tinyinference-llm/src/providers/anthropic/mod_tests.rs
• crates/tinyinference-llm/src/providers/anthropic/request.rs
• crates/tinyinference-llm/src/providers/media.rs
• crates/tinyinference-llm/src/providers/media_tests.rs
• crates/tinyinference-llm/src/providers/mod.rs
• crates/tinyinference-llm/src/providers/openai/convert.rs
• crates/tinyinference-llm/src/providers/openai/mod_tests.rs
• crates/tinyinference-llm/src/providers/openai/responses.rs
• crates/tinyinference-llm/src/providers/openai/responses_tests.rs
• crates/tinyinference-llm/src/providers/openai/transport.rs
• crates/tinyinference-llm/src/providers/openai/wire_tests.rs

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-04T06:13:06.721604Z	84d4caf	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[tinysweeper]

Requesting changes: 1 lane(s) blocking, worst finding is high.

Fix or reply to the findings below and push. The next review clears this automatically once they are gone — you should not need to dismiss anything by hand.

             $0.0011 · 51,124 in / 4,615 out · 25,856 cached (51%) · deepseek/deepseek-v4-flash
description: $0.0007 · 25,075 in / 341 out   · 0 cached (0%)       · deepseek/deepseek-v4-flash

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c76ed9eac0

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[senamakel]

Fixed the socket-fixture finding in f57b99a. The newly added successful-media tests now inspect synthetic provider request payloads and rely on synthetic byte-stream reconstruction coverage; the TCP listener fixture was removed. Public invoke/stream malformed-media rejection tests still assert a validation error before transport can run. Assistant text, thinking omission, tool IDs and serialized arguments remain covered by a pure serializer regression.

Fresh verification: cargo test --workspace --all-features, cargo clippy --workspace --all-targets --all-features -- -D warnings, and cargo fmt --all -- --check all pass. The PR description now lists these exact commands.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f57b99adba

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[senamakel]

Independent code review (Luna Agents), PR #61, exact head 84d4caf: APPROVED. No critical or important findings. The review verified the capability remains default-closed, unknown/denied formats are intersected safely, unsupported/local-path media is rejected before I/O, Responses PDF support remains opt-in and defaults off, and wrapper forwarding is correct. CI checks for stable Rust, Rust 1.88 MSRV, and supply chain are green. TinySweeper timed out with an explicit report that no code was reviewed; its check instructions permit hand merge after independent review.