Register telegram_send_video tool

[konard]

Summary

• Add the missing core telegram_send_video media tool and register it with the Telegram media toolset.
• Send local videos through GramJS as streamable Telegram videos with optional caption, reply target, duration, width, and height.
• Add a regression test proving registerAllTools exposes telegram_send_video.

Fixes #208

Reproduction

Before this change, calling telegram_send_video returned Unknown tool: telegram_send_video because the tool was documented and listed in send-tool constants but absent from the runtime media registration.

Verification

• npm test -- --run src/agent/tools/__tests__/registry.test.ts src/agent/tools/telegram/media/__tests__/send-voice.test.ts
• npm test -- --run
• npm run lint
• npm run build:sdk

Note: npm run typecheck currently fails locally because root TypeScript cannot resolve the workspace package alias @teleton-agent/sdk; npm run build:sdk succeeds and this appears unrelated to the video tool change.

[konard]

Solution summary

Implemented and pushed the fix for issue 208.

PR: #209
Commit: f9851d2 Register Telegram video send tool

What changed:

• Added telegram_send_video as a real runtime tool in src/agent/tools/telegram/media/send-video.ts.
• Registered and exported it from the Telegram media toolset.
• It validates workspace paths and sends videos through GramJS with DocumentAttributeVideo, forceDocument: false, and streaming support.
• Added a regression test proving registerAllTools() exposes telegram_send_video.

Verification completed:

• npm test -- --run src/agent/tools/__tests__/registry.test.ts src/agent/tools/telegram/media/__tests__/send-voice.test.ts passed.
• npm test -- --run passed: 110 files, 2672 tests.
• npm run lint passed.
• npm run build:sdk passed.
• npm run typecheck was attempted, but fails locally on unresolved @teleton-agent/sdk workspace alias errors unrelated to this change.

PR 209 has been updated with the real title/body and marked ready for review. No GitHub Actions checks are currently reported for the branch in either the fork or upstream PR view.

---

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-5.4
• Provider: OpenAI
• Public pricing estimate: unknown
• Token usage: 54,739 input, 8,348 output, 1,984,128 cache read

📊 Context and tokens usage:

Total: (54.7K + 2.0M cached) input tokens, 8.3K output tokens

🤖 Models used:

• Tool: OpenAI Codex
• Requested: gpt-5.4
• Model: GPT-5.4 (gpt-5.4)

📎 Log file uploaded as Gist (11255KB)

• View complete solution draft log

---

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