fix: stabilize CI runner compatibility and governance docs

[vaur94]

Summary

• fix chat pipeline orchestration so stages reuse the real session pipeline id
• remove machine-specific test path assumptions that broke GitHub runners
• align the PR guide with the protected-branch policy now enforced on main

Verification

• bun run ci:verify

[Copilot]

Pull request overview

This PR stabilizes CI behavior and documentation by fixing pipeline orchestration to reuse the persisted pipeline ID, removing machine-specific filesystem assumptions in tests, and aligning PR governance docs with current branch protection expectations.

Changes:

• Ensure chat pipeline stages reuse the persisted pipeline_runs.id for a session (instead of a transient UUID).
• Replace hardcoded local paths in tests with repo-relative or temp-directory paths for runner compatibility.
• Update EN/TR PR guide docs to reflect protected-branch policies on main.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tests/unit/github/git-utils.test.ts	Makes the “current project” path repo-relative to work on CI runners.
tests/helpers/test-utils.ts	Uses real temporary directories instead of hardcoded /tmp/test-project.
src/index.ts	Synchronizes ctx.pipelineId from the persisted session pipeline each loop iteration.
docs/tr/pr-kilavuzu.md	Updates TR governance bullets to match current main protections.
docs/en/pull-request-guide.md	Updates EN governance bullets to match current main protections.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

createTempProjectDirectory() creates real temp directories but nothing in this helper or the call sites cleans them up, so repeated unit test runs will leak /tmp/opencode-ceo-test-* directories (and possibly .ceo/state.db files) on developer machines/CI runners. Consider tracking created directories and exposing a cleanup helper (to be called from test afterEach/afterAll), or allow callers to pass in a directory so tests that only need client don’t create filesystem state unnecessarily.

Suggested change

	import { mkdtempSync } from "node:fs"
	import { tmpdir } from "node:os"
	import { join } from "node:path"
	import { initializeDatabase } from "../../src/state/database.js"
	function createTempProjectDirectory(): string {
	return mkdtempSync(join(tmpdir(), "opencode-ceo-test-"))
	}
	import { mkdtempSync, rmSync } from "node:fs"
	import { tmpdir } from "node:os"
	import { join } from "node:path"
	import { initializeDatabase } from "../../src/state/database.js"
	const tempProjectDirectories: string[] = []
	function createTempProjectDirectory(): string {
	const directory = mkdtempSync(join(tmpdir(), "opencode-ceo-test-"))
	tempProjectDirectories.push(directory)
	return directory
	}
	export function cleanupTempProjectDirectories(): void {
	for (const directory of tempProjectDirectories) {
	try {
	rmSync(directory, { recursive: true, force: true })
	} catch {
	// Ignore errors during cleanup to avoid failing tests on best-effort deletion.
	}
	}
	tempProjectDirectories.length = 0
	}