fix(compression): pass custom providers to feasibility check

[LeonSGP43]

What does this PR do?

Passes custom_providers through _check_compression_model_feasibility() so auxiliary compression models inherit per-model context_length overrides the same way the main model path already does. This avoids false 256K startup warnings when the compression model resolves through a custom provider entry with a larger configured context window.

Related Issue

Fixes #23779

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

• Forwarded custom_providers into get_model_context_length() inside run_agent.py::_check_compression_model_feasibility()
• Reused live config loading in the feasibility check so auxiliary custom-provider overrides stay aligned with the main-model and switch-model paths
• Added a regression test proving the feasibility check forwards custom provider metadata for the auxiliary compression model
• Updated existing feasibility assertions to match the helper's current empty-list fallback shape

How to Test

1. Run uv run --frozen pytest -q -o addopts='' tests/run_agent/test_compression_feasibility.py
2. Run uv run --frozen ruff check run_agent.py tests/run_agent/test_compression_feasibility.py
3. Confirm test_feasibility_check_passes_custom_providers asserts the auxiliary feasibility lookup receives the configured custom_providers list

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: macOS 15.5

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

Screenshots / Logs

• 17 passed in 1.01s from tests/run_agent/test_compression_feasibility.py
• All checks passed! from ruff check run_agent.py tests/run_agent/test_compression_feasibility.py

[alt-glitch]

Duplicate of #19553. Same fix: pass custom_providers to compression feasibility check. Multiple competing PRs exist (#19553, #20724, #21953, #23563, #23789).

[teknium1]

This appears to be implemented on current main already by a separate duplicate fix. This is an automated hermes-sweeper review.

Evidence:

• agent/conversation_compression.py:136-146 now calls get_model_context_length(...) for the auxiliary compression model with custom_providers=agent._custom_providers.
• The behavior was added by 7becb19ea00c13bdff6f78b71aa3ddfb0bdb5378 (fix(auxiliary): forward custom_providers to compression model context-length detection).
• git tag --contains 7becb19ea shows the fix is included starting with v2026.5.16.

Thanks for the PR. The duplicate note in the discussion was accurate; the requested custom-provider forwarding path is now present on main.