docs: make browser testing + telemetry a governed process

Added browser testing and telemetry requirements to CLAUDE.md and
copilot-instructions.md. MANDATORY for all agents: run
scripts/browser_test.py before committing any UI change. Read
telemetry logs to diagnose issues.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: nihilistau

.github/copilot-instructions.md

@@ -206,6 +206,21 @@ python scripts/smart_test.py --smoke    # standalone script, same engine
 python scripts/smart_test.py --list     # dry-run: show what would run
 ```
 
+## Browser Testing & Telemetry (v1.43.0)
+
+**MANDATORY**: After ANY change to JS, CSS, HTML, or Jinja2 templates, run:
+```powershell
+python scripts/browser_test.py                    # Full Playwright test (39+ checks)
+python scripts/browser_test.py --report           # Read telemetry without testing
+```
+
+Never commit UI changes without a passing browser test. The test suite clicks every HUD panel, popup, overlay, keyboard shortcut, and district card in a real headless browser.
+
+**Telemetry**: `cosysim-telemetry.js` auto-captures clicks, errors, hotkeys, network failures on every scene page. Events POST to `/api/telemetry` and persist to `data/structured_logs.jsonl` via StructuredLogger. Check telemetry when diagnosing issues:
+```powershell
+python -c "import json; [print(json.loads(l).get('msg','')[:100]) for l in open('data/structured_logs.jsonl') if 'browser' in l]"
+```
+
 ## Code Versioning & Comment Standards (v1.42.1)
 
 All code files created or significantly modified MUST include versioning metadata and structured comments. This is a **standing rule** for all agents (Copilot, Claude Code, local LLM agents).

CLAUDE.md

@@ -146,6 +146,8 @@ Always use `get_config().get("dot.path", default)`. Never hardcode values.
 - **File naming**: `test_{module_name}.py` → `test_{behavior}()`
 - **Seeded characters**: lola, viktor, aria, frankie, mira are always present in DB fixtures.
 - Ignore `tests/test_agent_loop.py` and `tests/live_wire_test.py` (require live services).
+- **Browser testing**: After ANY JS/CSS/HTML change, run `python scripts/browser_test.py` (Playwright). Never commit UI changes without a passing browser test. Read telemetry: `python scripts/browser_test.py --report`
+- **Telemetry**: `cosysim-telemetry.js` captures all browser clicks, errors, hotkeys → `POST /api/telemetry` → `data/structured_logs.jsonl`. Always check telemetry after user reports issues.
 
 ## Frontend