fix(#4060): extend backoff gate to remote MCP HTTP errors

[aheritier]

🤖 Automated implementer agent — this comment was posted by the implementer bot from Docker Agentic Platform, not by a human developer

Stacked on PR #4062. Base retargets to main once that merges.

Refs #4060 (partial — A2A pacing and LSP crash-loop pacing deferred; see below)

What

Remote MCP servers responding with 503/429/5xx during the initialize handshake previously triggered a fresh connect attempt on every agent turn (the #4060 burst pattern). This PR fixes that by wrapping the HTTP status from the remote server in a *modelerrors.StatusError so the StartableToolSet backoff gate arming logic can pace retries.

Design

Wrap point — enrichConnectError in pkg/tools/mcp/remote.go. The oauthTransport already records the last HTTP error status via logErrorResponse (for any >= 400 response). enrichConnectError previously embedded that status only as text in the error message. It now also calls modelerrors.WrapHTTPError(status, nil, enriched), surfacing the status as a *StatusError in the chain.

What arms the gate. startBackoffRetryable checks for a *modelerrors.StatusError with a retryable HTTP status (429/408/5xx) via errors.As — exactly as it already does for RAG embedding failures. No regex heuristics; no new classification logic in the gate itself.

What does NOT arm (unchanged policy):

• Local stdio MCP spawn failures (missing binary, connection refused) — never reach enrichConnectError.
• 4xx client errors (400/401/403) — wrapped in *StatusError for structured access but RetryableHTTPStatus returns false → fail promptly.
• Auth paths (oauthDeclined, authorizationRequired) — handled by their own early-return paths before the status branch; unaffected.
• lifecycle.ErrServerUnavailable, ErrTransport, ErrAuthRequired, ErrInitTimeout, ErrSessionMissing — the gate classifier explicitly excludes all of these.

Deferred:

• A2A pacing: the agent-card resolver does not cleanly expose HTTP status in its error chain; deferred to a follow-up.
• LSP crash-loop pacing: lifecycle.ErrServerCrashed is produced only inside lspSession.Wait() which flows to the supervisor's internal watcher, not to supervisor.Start(). The gate never sees it via the current error propagation path; deferred.

Changed files

File	What changed
pkg/tools/mcp/remote.go	enrichConnectError: wrap HTTP-status branch with modelerrors.WrapHTTPError
pkg/tools/startable_backoff.go	Updated classifier doc comment: complete excluded-sentinel list (adds ErrInitTimeout, ErrSessionMissing), note deferred LSP crash-loop path
pkg/tools/mcp/remote_test.go	3 new tests: 503→retryable *StatusError; 403→non-retryable *StatusError; network failure→no *StatusError
pkg/tools/startable_backoff_test.go	4 new tests: lifecycle sentinels (ErrServerUnavailable, ErrTransport, ErrAuthRequired) do NOT arm gate; 4xx StatusError does NOT arm gate
docs/tools/mcp/index.md	Updated startup-failure note: remote MCP 5xx/429 now paced
docs/tools/lsp/index.md	Updated startup-failure note: crash-loop pacing not yet in place; supervisor policy is the throttle