fix(perf): bound sharp screenshot memory in hosted mode to prevent pod OOM

[SavioBS629]

What & why

fetchAutomationScreenshots decodes/compresses up to 5 session screenshots
concurrently (Promise.all) through sharp, and sharp was never
configured — so it ran with its default libvips cache and a thread pool
sized to CPU cores. Several tens-of-MB pixel buffers decoded at once, with
no cap on peak memory, is a real OOM risk on memory-constrained hosted pods
(a few concurrent calls can exceed the pod's headroom).

Change (scoped to hosted REMOTE_MCP mode; local unchanged)

• sharp.cache(false) + sharp.concurrency(1) — no libvips cache, single
  worker thread — when running hosted.
• Process screenshots sequentially when hosted, so peak memory is a single
  image rather than all of them at once.
• Local (npx) behaviour is unchanged — still the concurrent path with
  sharp defaults (local machines aren't memory-constrained).

Test

• tsc, eslint: clean
• Full suite: 72 files / 1118 tests pass
• tests/setup.ts sharp mock extended with cache/concurrency no-ops

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Central YAML (base), Organization UI (inherited), Workspace UI (inherited)

Review profile: ASSERTIVE

Plan: Enterprise

Run ID: 202b0be9-10ce-49f2-ba91-a46a28c3e657

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}