fix(executions): mark_verified actually sends valid+reason; add replay()

[mikemolinet]

Summary

Two changes, both in ExecutionsResource:

(1) Bug fix — mark_verified silently dropped both kwargs

The prior implementation accepted valid and reason as keyword args but always sent json={}:

def mark_verified(self, execution_id, *, valid=True, reason=None):
    return self._client._post(f\"/v1/executions/{execution_id}/verify\", json={})

The server treats absent body as valid=true, so the default-arg path produced the right outcome by accident. But every caller passing valid=False or reason=\"...\" got verified_success instead of their intent, silently. No exception, no warning, no test caught it.

The fix builds the body explicitly:

body = {\"valid\": valid}
if reason is not None:
    body[\"reason\"] = reason
return self._client._post(f\"/v1/executions/{execution_id}/verify\", json=body)

Pinned by 4 regression tests so a future refactor can't regress to the bug:

• Default-arg sends {\"valid\": True} (not {})
• valid=False lands in body
• reason=\"...\" lands in body
• reason=None is omitted (not serialized as null)

(2) New: client.executions.replay(execution_id)

Closes one of the endpoints_missing entries from cueapi-python #24's parity manifest. Server-side POST /v1/executions/{id}/replay already shipped on prod; this is pure SDK catch-up.

Returns the server's response dict unchanged (new execution_id, scheduled_for, status, triggered_by=\"replay\", replayed_from).

Tests

5 new (12 → 17 total). All pass.

No hosted-PR dependency

Both server-side endpoints already shipped. Pure SDK fix + catch-up.

Backlog row

Will be added per the parity discipline. The followup PR (workers + usage resources) ships in a separate branch to keep this PR's scope tight on the bug fix + single endpoint add.

How the bug was found

Surfaced while surveying cueapi-python's surface for parity work — comparing the SDK's mark_verified against the cueapi-cli cueapi executions verify command I just shipped (cueapi-cli #31). The CLI explicitly tested --valid / --invalid / --reason branches; comparing back to the SDK exposed that the SDK had silently been broken on those paths.

🤖 Generated with Claude Code

[govindkavaturi-art]

Genuine bug fix on mark_verified — the previous version sent json={} and ignored the valid/reason params, so callers couldn't actually mark verifications failed. Plus replay() addition. Approve.