fix: agent time out issue

[hchenxa]

What & why

#302

Type of change

• Bug fix
• New feature
• Refactor / cleanup
• Docs
• Other:

Testing

GOCACHE=/private/tmp/san-go-build-cache go test ./...

Checklist

• Follows Conventional Commits (feat:, fix:, docs:, …)
• All commits signed off (git commit -s) — certifies the DCO
• Tests pass locally
• Docs updated (if behavior or config changed)
• Read and follow the Code of Conduct

[san-ci]

[REVIEW NOTIFIER] This PR requires a review from @yanmxa.

Once this PR has been reviewed and has the lgtm label, it will be automatically squash-merged.

Suggested reviewer

Based on OWNERS, @yanmxa is suggested to review this PR.

How to get your PR reviewed:

• Address reviewer feedback and push changes
• Reviewer adds lgtm label when satisfied
• An approving review + lgtm will trigger auto-merge

[hchenxa]

/assign @yanmxa

[san-ci]

@hchenxa assigned @yanmxa to this PR.

[yanmxa]

Two concerns:

1. Default 60s → 180s weakens an intentionally tight signal. FirstChunkTimeout = 5m already handles reasoning models thinking before the first token; once streaming, both DeepSeek and Claude extended thinking emit reasoning as chunks, so 60s of silence is usually a real stall. #302 has no repro. Prefer keeping the default and letting operators tune per-provider via the new fields.

2. New StreamFirstChunkTimeout / StreamIdleTimeout on BuildParams have no callers — no config/flag sets them. Wire them up in this PR or drop until wired.

Nit: internal/core/agent.go:119 still says default (60s) — stale after this change.

[hchenxa]

@yanmxa thanks for your comments, it make sense, i'd like to revert the default value changes and will introduce a new setting for streaming timeout

[hchenxa]

@yanmxa please take a look at again