Add coordinate-space mapping (model grid <-> physical pixels)

[JE-Chen]

Computer-use agent batch (from multi-agent web research; verified absent). Full layers + tests + EN/Zh v45 docs + README.

Feature (utils/coordinate_space, pure-arithmetic + Pillow)

• CoordinateSpace maps a model grid <-> physical pixels both ways (round + clamp). normalized_space (Gemini-style square 1000x1000), xga_space (Anthropic-style downscale, aspect-preserved, never upscales), downscale_png (resize a screenshot to the model's input size via Pillow — already core).
• Why: every 2026 computer-use/VLA model clicks in a fixed virtual grid, not physical pixels — the repo had no observe/act transform for this (only a replay-time scale_coordinates multiplier). Closes that gap so agent backends can right-size screenshots and translate clicks back.
• Executor AC_to_physical / AC_to_model; MCP ac_*; Builder under Agent.

Verification

• 9 tests pass (normalized round-trip, grid scaling + clamp, XGA aspect/no-upscale, clamping bounds, PNG downscale under importorskip, executor round-trip, wiring, facade); ruff clean; radon no CC≥C; bandit clean; PySide6-free.

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

🟢 Metrics 24 complexity · 0 duplication

Metric	Results
Complexity	24
Duplication	0

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud