Cache generated docs in local circuit prompt

[694410194]

Summary

• include the generated tscircuit docs from https://docs.tscircuit.com/ai.txt in the local circuit prompt
• load the required props docs and optional generated docs in parallel
• cache generated docs during the process so benchmark runs do not refetch them for every prompt build
• make OpenAI-backed tests opt-in via RUN_OPENAI_TESTS to keep default test runs deterministic

Tests

• bun test tests/prompt-templates/create-local-circuit-prompt.test.ts --timeout 50000
• bun test --timeout 50000
• bun x tsc --noEmit
• bun run build

/claim #45

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 04261fab37

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Avoid caching empty docs after transient fetch failures

The new cache assignment generatedDocsCache ??= await fetchOptionalFileContent(...) permanently stores "" when https://docs.tscircuit.com/ai.txt is temporarily unavailable (network error/503). Because "" is not nullish, all later prompt builds skip refetching and never recover generated docs until process restart, so long-running sessions can miss the “most up-to-date tscircuit API reference” even after the endpoint recovers.

Useful? React with 👍 / 👎.