fix(provider): cache all system messages instead of only first 2

[zouhua021]

Problem

applyCaching() in provider/transform.ts slices system messages to only the first 2:

const system = msgs.filter((msg) => msg.role === "system").slice(0, 2)

This means additional system messages (from plugins via experimental.chat.system.transform, MCP instructions, AGENTS.md content, etc.) are never marked with cacheControl. Since the system prompt is the most stable prefix across turns, not caching the full prefix causes unnecessary cache misses and higher input token costs.

Change

Remove the .slice(0, 2) limit so ALL system messages get cache markers:

const system = msgs.filter((msg) => msg.role === "system")

Impact

• 1 line changed
• Maximizes prefix cache reuse for Anthropic, Bedrock, OpenRouter, and other providers that support prompt caching
• The system prompt is identical across turns within a session, so caching the full prefix is always beneficial
• No functional change to model behavior

[github-actions]

This PR doesn't fully meet our contributing guidelines and PR template.

What needs to be fixed:

• PR description is missing required template sections. Please use the PR template.

Please edit this PR description to address the above within 2 hours, or it will be automatically closed.

If you believe this was flagged incorrectly, please let a maintainer know.

[github-actions]

Thanks for your contribution!

This PR doesn't have a linked issue. All PRs must reference an existing issue.

Please:

1. Open an issue describing the bug/feature (if one doesn't exist)
2. Add Fixes #<number> or Closes #<number> to this PR description

See CONTRIBUTING.md for details.

[github-actions]

The following comment was made by an LLM, it may be inaccurate:

Based on my search, I found a related PR that may be worth reviewing:

Related PR:

• PR fix(provider): improve prompt caching #35982: fix(provider): improve prompt caching — fix(provider): improve prompt caching #35982

This PR also addresses prompt caching optimization in the provider. It may have already tackled similar caching improvements or could provide context on the caching strategy. Worth checking if it covers the same issue or complements this fix.

No other exact duplicates addressing the .slice(0, 2) system message limitation were found.

[github-actions]

This pull request has been automatically closed because it was not updated to meet our contributing guidelines within the 2-hour window.

Feel free to open a new pull request that follows our guidelines.