refactor: hard-fork GSD -> MSD (Make Software Done)

Mechanical rename produced by scripts/msd-rename.cjs: gsd/Gsd/GSD -> msd/Msd/MSD
across contents and paths, upstream package/repo coordinates -> @golem15/msd-core
and golem15com/msd-core. Deep links into upstream history, sibling upstream
packages, the GSD-2 import feature, CHANGELOG.md and .changeset/ are kept as-is.

Hand edits on top: MSD block-letter banner and logos, LICENSE copyright line,
package/plugin identity, regenerated lockfile, install-tree fixtures, derived
registries and benchmark baseline; migration checksum baseline re-locked
(MSD keeps its own install state, so no install had applied the old sums);
sort-order and regex-escaped expectations in tests adjusted.

scripts/benchmark-compact-content.cjs

@@ -9,7 +9,7 @@
  * what `workflow.compact_content=false` pays today, since an opted-out project's
  * spine reads every detail part back in) against the "on" token count (spine
  * alone — what `workflow.compact_content=true` pays, since the detail `Read`
- * never fires). See `.gsd/phase/enhance-4404-token-benchmark/40-design.md` for
+ * never fires). See `.msd/phase/enhance-4404-token-benchmark/40-design.md` for
  * the full design rationale.
  *
  * PROXY-TOKENIZER CAVEAT: Anthropic publishes no tokenizer for Claude 3+, so
@@ -63,7 +63,7 @@ const { countTokens } = require('gpt-tokenizer');
 const { runMain } = require('./lib/cli-exit.cjs');
 
 const ROOT = path.resolve(__dirname, '..');
-const WORKFLOWS_DIR = path.join(ROOT, 'gsd-core', 'workflows');
+const WORKFLOWS_DIR = path.join(ROOT, 'msd-core', 'workflows');
 const BASELINE_PATH = path.join(ROOT, 'tests', 'fixtures', 'compact-content-benchmark-baseline.json');
 
 // The tokenizer's own package.json is read live (`require.resolve` + a plain