Enable direct Web Search for AI chat

[konard]

Summary

• Expose a new run_web_search tool to AI chat for current rates, news, fiat/forex pairs, and other non-tracked topics.
• Route the chat tool through the existing Tradefast.search() server-side web search path with a bounded result limit.
• Add prompt guidance and README documentation so the assistant searches before refusing unsupported pairs.

Reproduction

• In normal AI chat, ask for a current non-tracked rate such as rub/usd rate.
• Before this change, the chat request did not advertise any direct web-search tool, so the assistant could only answer from static Tradefast context or ask the user to use configured crypto analysis.

Tests

• npm test -- tests/cli.test.tsx -t "lets AI chat" (fails before the fix, passes after)
• npm run typecheck
• npm test
• npm run build

Fixes #35

[konard]

Working session summary

Implemented and pushed the fix.

PR: #36

What changed:

• Added run_web_search as an AI chat tool.
• Wired it to existing Tradefast.search() so normal chat can search the internet directly.
• Updated the Russian system prompt to search before refusing fiat/forex/current-info questions.
• Added a regression test for rub/usd rate style chat.
• Updated README AI chat tools docs.
• Removed the placeholder .gitkeep.

Verification:

• npm test -- tests/cli.test.tsx -t "lets AI chat" passed
• npm run typecheck passed
• npm test passed: 147 tests
• npm run build passed
• GitHub CI build-and-test passed on commit df179c3

PR #36 is marked ready for review, and the working tree is clean.

---

This summary was automatically extracted from the AI working session output.

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Model: GPT-5.5
• Provider: OpenAI
• Public pricing estimate: $5.531475

📊 Context and tokens usage:

• 125.5K / 1.1M (12%) input tokens, 17.7K / 128K (14%) output tokens

Total: (125.5K + 3.5M cached) input tokens, 17.7K output tokens, $5.531475 cost

🤖 Models used:

• Tool: OpenAI Codex
• Requested: gpt-5.5
• Model: GPT-5.5 (gpt-5.5)

📎 Log file uploaded as Gist (23886KB)

• View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.