Specificity Gap

Instructions must name concrete constructs -- backtick-wrapped tokens, file paths, function names, CLI commands -- instead of abstract concepts. Abstract instructions are dramatically less effective because the model cannot distinguish them from general knowledge.

Antipatterns

• Writing "Follow the coding style" instead of naming the specific tool (ruff format, 4-space indent, snake_case). The model interprets abstract style references using its own defaults.
• Using category names like "mocking libraries" instead of specific imports like unittest.mock, MagicMock, patch(). Category names are as vague as abstract concepts.
• Stating "Run the tests" without specifying the command (uv run pytest tests/ -v). The model guesses which test runner to use.

Pass / Fail

Pass

Use `ruff format` with 4-space indent and `snake_case`
for all functions in `src/reporails_cli/`.
Run `uv run pytest tests/ -v` before committing.

Fail

Follow the project's coding style.
Run the tests before committing.
Use appropriate mocking libraries.

Fix

Replace "Don't use mocking" with "Don't use unittest.mock, MagicMock, patch()". Replace "Follow the coding style" with "Use ruff format, 4-space indent, snake_case for functions". Name the exact tools, functions, files, patterns, and libraries. Category names ("mocking libraries") are as vague as abstract concepts — the model needs the import path, not the category.

Limitations

Measures whether instructions contain named constructs (backtick-wrapped tokens, file paths, function names). Cannot evaluate whether the named constructs are the right ones for the project.