Add Phase 5 & Phase 6 test coverage: Session/Agent and Server Routes

[Copilot]

Implements test coverage for Phase 5 (Session/Agent) and Phase 6 (Server Routes) from TEST_COVERAGE_PLAN.md, with fixture-driven test data.

Phase 5 — Session/Agent (test/session/)

• system.test.ts — SystemPrompt.provider() routing (all 6 branches: gpt-5/codex, gpt-*/o1/o3/beast, gemini-, claude/anthropic, trinity, fallback), instructions(), environment() with git/non-git repos
• summary.test.ts — SessionSummary.computeDiff() (empty/missing-snapshot cases), SessionSummary.diff() including unquoteGitPath via Storage round-trips with git-quoted octal-escaped paths
• processor.test.ts — SessionProcessor.create() structure: message getter identity, partFromToolCall() unknown-ID lookup, independent state across instances, abort signal lifecycle
• Fixtures (test/session/fixtures/): messages-with-snapshots.json, session-diff-quoted.json

Phase 6 — Server Routes (test/server/)

All tests use Server.Default() + app.request() with x-opencode-directory header — consistent with existing server test pattern.

• error.test.ts — pure unit: ERRORS object shape, errors() with 0/1/N codes
• config.test.ts — GET /config, PATCH /config (valid + invalid payload), GET /config/providers
• provider.test.ts — GET /provider (all/default/connected fields, connected ⊆ all), GET /provider/auth
• mcp.test.ts — GET /mcp empty status, disconnect idempotency
• file.test.ts — /find/file with limit, /file/status git untracked file, /file listing, /find/symbol (LSP disabled → [])
• permission.test.ts — GET /permission empty + populated via PermissionNext.ask(), POST once/reject reply flows
• question.test.ts — GET /question empty + populated via Question.ask(), POST reply/reject flows
• experimental.test.ts — /experimental/tool/ids (includes bash/read/write), /experimental/tool validation, /experimental/session pagination + directory filter
• Fixtures (test/server/fixtures/): config-patch.json, file-find-query.json

Type of change

• Bug fix
• New feature
• Refactor / code improvement
• Documentation

What does this PR do?

Covers the 11 previously untested source modules listed in TEST_COVERAGE_PLAN.md phases 5–6. Each test exercises the real implementation (no mocks except where I/O is unavoidable). File-based JSON fixtures drive parameterized assertions — e.g. git-quoted path decoding, config patch fields, file search limits.

session/processor.ts full process() loop and server/routes/pty.ts/tui.ts remain excluded per plan (require live AI stream and TUI terminal respectively).

How did you verify your code works?

Tests follow established patterns (Instance.provide + tmpdir, Session.create/remove, Storage.write/read, Server.Default() + app.request()). CodeQL scan returned zero alerts.

Screenshots / recordings

N/A — no UI changes.

Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR

---

📍 Connect Copilot coding agent with Jira, Azure Boards or Linear to delegate work to Copilot in one click without leaving your project management tool.

[github-actions]

Hey! Your PR title Add Phase 5 Session/Agent test coverage with fixture data doesn't follow conventional commit format.

Please update it to start with one of:

• feat: or feat(scope): new feature
• fix: or fix(scope): bug fix
• docs: or docs(scope): documentation changes
• chore: or chore(scope): maintenance tasks
• refactor: or refactor(scope): code refactoring
• test: or test(scope): adding or updating tests

Where scope is the package name (e.g., app, desktop, opencode).

See CONTRIBUTING.md for details.

[github-actions]

This PR doesn't fully meet our contributing guidelines and PR template.

What needs to be fixed:

• PR description is missing required template sections. Please use the PR template.

Please edit this PR description to address the above within 2 hours, or it will be automatically closed.

If you believe this was flagged incorrectly, please let a maintainer know.

[lknight]

Больше тестов перед началом изменений