feat(go-sdk): deduplicate tool_calling, add ToolCallResult wrapper, PromptConfig, system prompt support, and tool name sanitization

[Luffy2208]

Summary

Refactors the Go SDK tool-calling loop to improve code quality, remove duplication, and align behavior with Python and TypeScript SDKs. This PR introduces structured abstractions, configurable prompts, system prompt support, and tool name sanitization while preserving backward compatibility.

---

Key Changes

1. Deduplicated Capability Conversion

• Removed duplicated logic for ReasonerCapability and SkillCapability
• Introduced shared helpers:
  • capabilityToTool(...)
  • normalizeToolParameters(...)
• Updated CapabilityToToolDefinition to delegate to helper functions

---

2. Added ToolCallResult Wrapper (API Parity)

• Introduced:

  type ToolCallResult struct {
      Response *Response
      Trace    *ToolCallTrace
  }

• Added .Text() helper for convenience
• Introduced ExecuteToolCallLoopResult(...)
• Preserved backward compatibility via existing ExecuteToolCallLoop(...)

---

3. Introduced PromptConfig (Removed Hardcoded Strings)

• Added:

  type PromptConfig struct {
      ToolCallLimitReached string
      ToolErrorFormatter   func(...)
      ToolResultFormatter  func(...)
  }

• Implemented:
  • DefaultPromptConfig()
  • resolvePromptConfig(...)
• Replaced all hardcoded tool-related messages with configurable formatters

---

4. Added System Prompt Support

• Extended ToolCallConfig:

  SystemPrompt string

• Injected using:

  WithSystem(config.SystemPrompt)

---

5. Added Tool Name Sanitization

• Implemented:
  • sanitizeToolName(...)
  • unsanitizeToolName(...)
• Applied across:
  • Tool definitions
  • Tool execution
  • Trace recording

---

6. Refactored Tool-Calling Loop

• Extracted request builder:
  • buildToolCallRequest(...)
• Introduced:
  • loopMessages
  • effectiveOpts
• Removed duplicated request-building logic
• Improved error handling flow

---

7. Added encodeToolContent Helper

• Centralized encoding logic for:
  • Tool errors
  • Tool results
  • Limit messages
• Handles string/JSON formatting safely

---

8. Improved Default Configuration

• Updated DefaultToolCallConfig() to initialize PromptConfig

---

9. Updated agent.go for Compatibility

• Added transformation layer:
  • Converts :-based tool names to . format before execution

---

10. Expanded Test Coverage

Added tests for:

• Tool name sanitization (round-trip + capability conversion)
• PromptConfig behavior (limit messages, error formatting, result formatting)
• System prompt injection
• ToolCallResult wrapper
• Default config initialization

---

Testing

• ./scripts/test-all.sh
• Ran go test ./... successfully
• Ran go vet ./... successfully

---

Checklist

• I updated documentation where applicable.
• I added or updated tests (or none were needed).
• I updated CHANGELOG.md (or this change does not warrant a changelog entry).

---

Screenshots (if UI-related)

Not applicable

---

Related Issues

Fixes #234

[github-actions]

Performance

SDK	Memory	Δ	Latency	Δ	Tests	Status
Go	210 B	-25%	0.57 µs	-43%	✓	✓

✓ No regressions detected

[github-actions]

📊 Coverage gate

Thresholds from .coverage-gate.toml: per-surface ≥ 86%, aggregate ≥ 88%, max per-surface regression ≤ 1.0 pp, max aggregate regression ≤ 0.50 pp.

Surface	Current	Baseline	Δ
control-plane	87.20%	87.30%	↓ -0.10 pp	🟡
sdk-go	90.70%	90.70%	→ +0.00 pp	🟢
sdk-python	93.63%	93.63%	↑ +0.00 pp	🟢
sdk-typescript	92.63%	92.56%	↑ +0.07 pp	🟢
web-ui	90.02%	90.01%	↑ +0.01 pp	🟢
aggregate	88.98%	89.01%	↓ -0.03 pp	🟡

✅ Gate passed

No surface regressed past the allowed threshold and the aggregate stayed above the floor.

[github-actions]

📐 Patch coverage gate

Threshold: 80% on lines this PR touches vs origin/main (from .coverage-gate.toml:thresholds.min_patch).

Surface	Touched lines	Patch coverage	Status
control-plane	0	—	➖ no changes
sdk-go	140	84.00%	✅
sdk-python	0	—	➖ no changes
sdk-typescript	0	—	➖ no changes
web-ui	0	—	➖ no changes

✅ Patch gate passed

Every surface whose lines were touched by this PR has patch coverage at or above the threshold.

[Luffy2208]

Working on test coverage