Fix compaction cascade logging for single-provider configs

[konard]

Fixes #275

Summary

• Resolve optional compaction cascade short names with the non-logging resolver so fallback misses do not emit provider-level error logs.
• Log unresolved default cascade entries at debug level while keeping explicit --compaction-models misses visible as warnings.
• Add a regression test for a single custom provider config matching the issue shape.

Reproduction

Using the issue's LINK_ASSISTANT_AGENT_CONFIG_CONTENT with only the formalai/formal-ai custom provider, verbose startup previously emitted repeated model not found - refusing to silently fallback error logs while probing the default compaction cascade. After this change, those default fallback misses are debug-level skip logs and no provider error burst is emitted.

Tests

• bun test ./tests/compaction-model.ts
• npm test
• npm run check
• cargo test --all-features

[konard]

Working session summary

Implemented and pushed the fix to PR 276: #276

Changed compaction cascade resolution so default fallback misses no longer use the strict provider resolver that logs ProviderModelNotFoundError at error level. Default cascade skips now log at debug; explicit --compaction-models skips still warn. Added a regression test for the single custom provider config from issue 275 and added the required patch changeset.

Verification passed:

• bun test ./tests/compaction-model.ts
• npm test
• npm run check
• cargo test --all-features
• node scripts/validate-changeset.mjs
• GitHub CI JS CI/CD Pipeline passed on head 0b8bc3b

PR is ready for review, merge state is clean, and the local working tree is clean.

---

This summary was automatically extracted from the AI working session output.

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Model: GPT-5.5
• Provider: OpenAI
• Public pricing estimate: $12.847237

📊 Context and tokens usage:

• 422.6K / 200K (211%) input tokens, 33.1K / 128K (26%) output tokens

Total: (422.6K + 7.1M cached) input tokens, 33.1K output tokens, $12.847237 cost

🤖 Models used:

• Tool: OpenAI Codex
• Requested: gpt-5.5
• Thinking level: max (~31999 tokens)
• Model: GPT-5.5 (gpt-5.5)

📎 Log file uploaded as Gist (4178KB)

• View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.

[konard]

🎉 Auto-merged

This pull request has been automatically merged by hive-mind.

• All CI checks have passed

---

Auto-merged by hive-mind with --auto-merge flag