Skip to content

CCL Prompts for AI Agents

Copy-paste these prompts to guide AI agents building CCL implementations.

  1. Choose a prompt that matches your goal
  2. Replace [language] with your target language (Rust, Python, etc.)
  3. Paste into your AI agent (Claude, ChatGPT, Copilot, etc.)
  4. The AI will read the docs and follow the specification

Build a CCL parser in [language].
Read https://ccl.tylerbutler.com/ai-implementation-guide/ for the complete
specification. Start with parse and build_hierarchy, then add typed access.
Use tests from https://github.com/tylerbutler/ccl-test-data to validate.
Build a CCL parser in [language] with these behaviors:
- crlf_normalize_to_lf (normalize line endings)
- boolean_strict (only "true" and "false")
- list_coercion_disabled (require explicit list syntax)
Read https://ccl.tylerbutler.com/ai-implementation-guide/ for the specification.
See https://ccl.tylerbutler.com/behavior-reference/ for behavior details.
Build a minimal CCL parser in [language] with only the required functions:
- parse (text to entries)
- build_hierarchy (entries to nested object)
Skip typed access and formatting for now.
Read https://ccl.tylerbutler.com/ai-implementation-guide/#core-functions-required

I have a CCL parser with parse and build_hierarchy working.
Add typed access functions:
- get_string
- get_int
- get_bool
- get_float
- get_list
Follow https://ccl.tylerbutler.com/ai-implementation-guide/#typed-access-functions-optional
Use https://ccl.tylerbutler.com/library-features/ for implementation details.
Add the get_list function to my CCL implementation following the spec at
https://ccl.tylerbutler.com/ai-implementation-guide/#get_list
Use list_coercion_disabled behavior (error if value is not actually a list).
Add formatting functions to my CCL implementation:
- print (structure-preserving round-trip)
- canonical_format (normalized output)
Follow https://ccl.tylerbutler.com/library-features/#formatting-functions
Add entry processing functions to my CCL implementation:
- filter (remove entries by predicate, e.g., comments)
- compose (concatenate entry lists)
Follow https://ccl.tylerbutler.com/library-features/

Set up test integration for my CCL implementation using the test suite at
https://github.com/tylerbutler/ccl-test-data
The tests are in generated_tests/*.json format. Filter tests to only run
those for functions I've implemented: parse, build_hierarchy.
See https://ccl.tylerbutler.com/test-suite-guide/ for test format details.
Build a test runner for my [language] CCL implementation that:
1. Loads JSON tests from https://github.com/tylerbutler/ccl-test-data/generated_tests/
2. Filters by my implemented functions: [list your functions]
3. Runs each test and reports pass/fail
4. Handles the "expected" format with "count" fields
See https://ccl.tylerbutler.com/test-suite-guide/ for the test format.
Run only the parse function tests from the CCL test suite.
Filter tests where validation === "parse" from the generated_tests/ directory.
Report which tests pass and fail.

My CCL parser is failing these tests: [paste test names or errors]
Review my implementation against the algorithm at
https://ccl.tylerbutler.com/parsing-algorithm/ and identify issues.
Common pitfalls are listed at
https://ccl.tylerbutler.com/ai-implementation-guide/#common-pitfalls
Review my CCL implementation for correctness.
Check against:
- Algorithm: https://ccl.tylerbutler.com/parsing-algorithm/
- Common pitfalls: https://ccl.tylerbutler.com/ai-implementation-guide/#common-pitfalls
- Behaviors: https://ccl.tylerbutler.com/behavior-reference/
Identify any deviations from the specification.
Review my CCL implementation's error handling.
The spec at https://ccl.tylerbutler.com/ai-implementation-guide/ describes
expected error cases. Ensure my implementation handles:
- Missing keys (path not found)
- Type mismatches (e.g., get_int on non-integer)
- Malformed input (lines without =)

Implement CCL in [language] following https://ccl.tylerbutler.com/ai-implementation-guide/
Read https://ccl.tylerbutler.com/ai-quickstart/ for CCL terminology and concepts.
Compare my CCL parser to the spec at https://ccl.tylerbutler.com/parsing-algorithm/

Be specific about functions: Instead of “add more features,” say “add get_string and get_int.”

Specify behaviors: If you care about boolean parsing or list coercion, say so explicitly.

Reference specific URLs: Point to exact sections like #get_list for focused work.

Include your constraints: Language version, dependencies, style preferences.

Mention the test suite: Always reference https://github.com/tylerbutler/ccl-test-data for validation.