Standardize configuration defaults and remove machine-specific naming

[KulmanD]

Objective

Establish one authoritative set of defaults and remove the machine-specific config/denis-config.json naming without modifying the user's existing configuration file.

Planned branch: refactor/standardize-config-defaults
Performance report: docs/performance/standardize-config-defaults.md
Performance gate: #20

Scope

• Rename the shared example/default configuration to a neutral path.
• Resolve the current maxScrollStep disagreement between the JSON file and compiled Swift defaults.
• Remove an unused bundled default resource or make its ownership explicit; do not retain two independent default sources.
• Update packaging, app-building, README, and technical documentation references.
• Preserve ~/Library/Application Support/mx3-lite/config.json exactly.
• Do not silently migrate or overwrite user settings.
• Keep missing required button IDs mapped to disabled behavior.

Tests

• first-run config creation
• existing config preservation
• partial config fallback
• invalid-value warnings
• app and CLI resolve identical defaults
• source and app packages contain the intended neutral file
• release artifacts exclude personal paths and data

Performance experiment

Although this should be runtime-neutral, verify:

• release binary and app archive sizes
• strict idle worker measurement
• startup/config-load time
• RSS and thread count
• no new resource parsing in the worker
• no test framework in release output

Definition of done

• One documented authority owns defaults.
• No machine-specific filename remains in distribution.
• Existing user config is untouched.
• Measured worker behavior remains equivalent to main.

[KulmanD]

Implementation and deterministic validation are complete on refactor/standardize-config-defaults (draft PR #36). The release helper and app executable are byte-size neutral in equal-path clean builds, parsed config output is identical, the existing user config remained byte-for-byte unchanged, packaging checks pass, and GitHub Actions previously passed 34 tests. Two 150-second idle runs recorded zero package/interrupt wakeup growth; the optional third repetition was deliberately deferred so the remaining issues can be completed. Full evidence: docs/performance/standardize-config-defaults.md.