mirror of
https://github.com/BerriAI/litellm.git
synced 2026-10-12 03:48:56 +00:00
e7a9a8138b
Adds tooling for an agent-driven loop that hardens a provider's e2e
test suite (coverage + mutation) before refactoring its source. Source
edits are blocked via chmod + a sentinel-gated pre-commit hook so tests
can't drift to cover internals the refactor will change.
- scripts/e2e_llm_loop.sh: per-provider pytest+coverage and mutmut runs
- scripts/e2e_llm_lock.sh: chmod-based source lock
- scripts/e2e_llm_install_hook.sh: belt-and-suspenders pre-commit hook
- Makefile: e2e-llm-{coverage,mutation,lock,unlock,hook-*} targets
- tests/llm_translation/E2E_AGENT.md: agent prompt and operating loop
- .gitignore: defensive blocks on file-based VCR cassettes
- pyproject.toml: branch coverage + report config
https://claude.ai/code/session_01MWq8irLUTyWi79JXY8TBHy
|
||
|---|---|---|
| .. | ||
| adaptive_router_demo | ||
| health_check | ||
| benchmark_mock.py | ||
| benchmark_proxy_vs_provider.py | ||
| create_litellm_branch.ps1 | ||
| create_litellm_branch.sh | ||
| create_team_key_and_submit_guardrail.sh | ||
| e2e_llm_install_hook.sh | ||
| e2e_llm_lock.sh | ||
| e2e_llm_loop.sh | ||
| eval_compression.py | ||
| install.sh | ||
| mock_bedrock_passthrough_target.py | ||
| mock_grayswan_timeout_server.py | ||
| mutation_report.py | ||
| test_agent_mcp_endpoints.sh | ||
| test_guardrails_register_endpoints.sh | ||
| test_tool_allowlist_script.py | ||
| tpm_headline_test.sh | ||
| verify_adaptive_router.py | ||