fix(apollo): guard plan reranker env provider

@acsrujanchecks n/achecks…fix/plan-reranker-unsupported-env-provider → master2 files · +73 −3review: @wjayeshupdated 3w ago

GitHub

▸Description · 6 noise

Description

Prevent plan-search reranking from silently parsing unsupported RERANKER_MODEL providers, such as bedrock:..., as Groq model names.
Keep the existing supported plan-search providers (groq, cerebras, openai, gemini) working through the existing parser.
Fall back to LaunchDarkly/default plan-search reranker config when the env provider is unsupported, and log the ignored provider.
Add regression coverage for the bedrock: case raised by Cursor Bugbot on the master-to-production release PR.

How did I test this PR

pnpm exec oxfmt --check apps/apollo/src/lib/composio_actions/utils/planSearch/clients.ts apps/apollo/src/lib/composio_actions/utils/planSearch/__tests__/clients.rerank.unit.test.ts
pnpm exec oxlint apps/apollo/src/lib/composio_actions/utils/planSearch/clients.ts apps/apollo/src/lib/composio_actions/utils/planSearch/__tests__/clients.rerank.unit.test.ts
git diff --check

Note

Focused Vitest is still blocked in this local checkout before test collection by missing package resolution: mint-id/extended imported from apps/apollo/src/lib/toolRouterV2/utils/wordId.ts.

▸ 6 bot/status comments hidden

@vercel[bot]3w ago

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
apollo	Ready	Preview, Comment	May 19, 2026 11:07am

Project	Deployment	Actions	Updated (UTC)
debby	Skipped		May 19, 2026 11:07am

@github-actions[bot]3w ago

🔍 Suggested Reviewers

Based on git blame analysis of 2 file(s):

Contributor	Contribution	Files
Zen Agent	63%	2
Dhawal Upadhyay	36%	1
Jayesh Sharma	1%	1
devin-ai-integration[bot]	0%	1
lingalarahul7	0%	1

💡 Recommendation

I suggest Zen Agent — they made recent (today) edits to both the test and implementation files and will be best placed to review the immediate changes. Also add Dhawal Upadhyay — primary author of clients.ts and familiar with the implementation details for a deeper logic/API review.

🤖 Based on git blame with recency weighting (recent edits count more).

@github-actions[bot]3w ago

🧪 Test Results Summary (Unit Tests)

@github-actions[bot]3w ago

🧪 Test Results Summary (Self-Hosted Tests)

@codecov[bot]3w ago

Codecov Report

:x: Patch coverage is 92.85714% with 2 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...c/lib/composio_actions/utils/planSearch/clients.ts	92.85%	2 Missing :warning:

Flag	Coverage Δ
e2e-tests	`6.05% <0.00%> (-0.01%)`	:arrow_down:
self-hosted-tests	`5.59% <0.00%> (-0.01%)`	:arrow_down:
thermos-unit-tests	`?`

@github-actions[bot]3w ago

🧪 Test Results Summary (E2E Tests)

loading diff…

Description

Prevent plan-search reranking from silently parsing unsupported RERANKER_MODEL providers, such as bedrock:..., as Groq model names.

Keep the existing supported plan-search providers (groq, cerebras, openai, gemini) working through the existing parser.

Fall back to LaunchDarkly/default plan-search reranker config when the env provider is unsupported, and log the ignored provider.

Add regression coverage for the bedrock: case raised by Cursor Bugbot on the master-to-production release PR.

How did I test this PR

pnpm exec oxfmt --check apps/apollo/src/lib/composio_actions/utils/planSearch/clients.ts apps/apollo/src/lib/composio_actions/utils/planSearch/__tests__/clients.rerank.unit.test.ts

pnpm exec oxlint apps/apollo/src/lib/composio_actions/utils/planSearch/clients.ts apps/apollo/src/lib/composio_actions/utils/planSearch/__tests__/clients.rerank.unit.test.ts

git diff --check