fix(execute): propagate agent 4xx in async lane instead of blanket 502

[7vignesh]

Summary

Fixes #862. The async-completion branch in handleSync hardcoded HTTP 502 for all failed executions, making client-input rejections (e.g. 422 from input validation) indistinguishable from upstream outages. A reasoner validating its own input would return invalid_input: ... but the caller saw 502 Bad Gateway — sending operators to check pods instead of the request.

Root cause: The callError.statusCode is lost once the execution is persisted; the async lane (which reads the execution from DB after the agent calls back) had no way to recover the original HTTP status.

Fix (two-pronged):

Control plane

• Add error_status_code field to executionStatusUpdateRequest so the SDK can forward the HTTP status in its failure callback
• When error_status_code is 4xx, encode it in StatusReason as agent_client_error:<code> for persistence
• Add httpStatusForFailedExecution() helper that resolves the HTTP status from:
  1. StatusReason — encoded client errors and known server-side categories (agent_timeout→504, target_not_found→404, etc.)
  2. ErrorMessage — parses "agent error (NNN):" pattern as fallback for existing executions
  3. Default: 502
• Replace hardcoded 502 in the async-completion branch with the helper

Python SDK

• In _execute_async_with_callback failure path, propagate error_status_code from exception.status_code / exception.code when the value is a valid HTTP status (400–599)

Backward-compatible: SDKs that don't send error_status_code get existing behavior (502 default). SDKs that do send it get correct 4xx propagation immediately.

Type of change

• Bug fix

Test plan

• cd control-plane && go test ./internal/handlers/ -run TestHttpStatusForFailedExecution -v (12 cases: encoded 4xx, category mapping, error message parsing, defaults)
• cd control-plane && go test ./internal/handlers/ -run TestUpdateExecutionStatusHandler_ErrorStatusCode -v
• cd control-plane && go test ./internal/handlers/ -run TestExecuteHandler_AgentError -v (existing 500→502 preserved)
• cd control-plane && go test ./internal/handlers/ -timeout 120s (full suite)
• cd control-plane && go vet ./internal/handlers/...
• cd control-plane && golangci-lint run ./internal/handlers/ (no new warnings)
• cd sdk/python && python -m pytest tests/test_usage_transport.py (async callback path)
• python -c "import agentfield.agent" (syntax OK)

Test coverage

• I ran tests for the surface(s) I changed locally.
• New code paths are covered by tests in this PR (no bare additions).
• If I removed code, I updated coverage-baseline.json in this PR only if the removal caused a legitimate regression and I called it out in the summary above.
• The coverage gate check is green in CI before requesting review.

Checklist

• Commits follow conventional-commits style.
• I have linked any related issues (Fixes execute: async lane returns 502 for a client-input rejection; sync lane returns the agent's 4xx #862).

Related issues / PRs

Fixes #862

Note: TestCallAgent_ErrorResponse is a pre-existing flaky test (timing-sensitive elapsed > 0 assertion) — confirmed by running it with -count=5 on both main and this branch.

[github-actions]

Performance

SDK	Memory	Δ	Latency	Δ	Tests	Status
Python	9.0 KB	-	0.33 µs	-6%	✓	✓

✓ No regressions detected

[github-actions]

📊 Coverage gate

Thresholds from .coverage-gate.toml: per-surface ≥ 84%, aggregate ≥ 85%, max per-surface regression ≤ 1.0 pp, max aggregate regression ≤ 0.50 pp.

Surface	Current	Baseline	Δ
control-plane	87.20%	87.40%	↓ -0.20 pp	🟡
sdk-go	92.90%	92.00%	↑ +0.90 pp	🟢
sdk-python	94.20%	93.73%	↑ +0.47 pp	🟢
sdk-typescript	91.39%	90.42%	↑ +0.97 pp	🟢
web-ui	84.76%	84.79%	↓ -0.03 pp	🟡
aggregate	85.66%	85.75%	↓ -0.09 pp	🟡

✅ Gate passed

No surface regressed past the allowed threshold and the aggregate stayed above the floor.

[github-actions]

📐 Patch coverage gate

Threshold: 80% on lines this PR touches vs origin/main (from .coverage-gate.toml:thresholds.min_patch).

Surface	Touched lines	Patch coverage	Status
control-plane	40	92.00%	✅
sdk-go	0	—	➖ no changes
sdk-python	0	—	➖ no changes
sdk-typescript	0	—	➖ no changes
web-ui	0	—	➖ no changes

✅ Patch gate passed

Every surface whose lines were touched by this PR has patch coverage at or above the threshold.

[santoshkumarradha]

Validated the control-plane and Python surfaces locally. The status propagation fix and regression coverage look good to me.