feat(codegen): replace generic <value> placeholders with type-aware placeholders

[pyramation]

Summary

Generated documentation and skill reference files previously used generic <value> placeholders in all code examples, which were non-descriptive. This PR replaces them with type-aware placeholders derived from the GraphQL schema (e.g. <UUID>, <String>, <Int>, <Boolean>).

Before: db.car.findOne({ id: '<value>' ... })
After: db.car.findOne({ id: '<UUID>' ... })

Three new helpers were added to docs-utils.ts:

• fieldPlaceholder(field) — for table fields (CleanField with type.gqlType)
• pkPlaceholder(pk) — for primary key fields (PrimaryKeyField with direct gqlType)
• argPlaceholder(arg, registry?) — for operation arguments; expands INPUT_OBJECT types into constituent fields when ≤5 fields, otherwise shows the type name

flattenedArgsToFlags() was also updated to use <TypeName> in CLI flag docs.

The TypeRegistry is now threaded through generate.ts → ORM/hooks doc generators so argPlaceholder can resolve INPUT_OBJECT fields.

Review & Testing Checklist for Human

• Verify argPlaceholder INPUT_OBJECT expansion: The function expands input objects with ≤5 fields (excluding clientMutationId). Run codegen against a schema with input types to confirm the sign-in example now shows { email: '<String>', password: '<String>' } instead of '<value>'
• Confirm pkPlaceholder vs fieldPlaceholder distinction is correct everywhere: PrimaryKeyField has gqlType at top level while CleanField nests it under .type.gqlType. Grep for pkPlaceholder and fieldPlaceholder usage to verify no mismatches
• Check ORM update examples: The update code examples changed from '<new-value>' (semantic hint) to the field's actual type placeholder — verify this tradeoff is acceptable
• Run full codegen against a real schema (e.g. the auth schema that produces sign-in.md) to verify the actual output matches expectations

Notes

• Semantic placeholders in CLI config/auth commands (<key>, <value>, <token>, <name>) are intentionally unchanged — they describe the semantic role, not a typed field
• The hardcoded '<UUID>' in generic SKILL.md examples (db.<model>.findOne({ id: '<UUID>' ...})) assumes UUID primary keys, which is the common case
• 6 test failures in CI are pre-existing module resolution issues (e.g. @pgpmjs/core, pg-cache) unrelated to this change

Link to Devin session: https://app.devin.ai/sessions/6acc9cfe2eb74b44819e6c7e3d747d68
Requested by: @pyramation

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring