Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Cpp Pro · davila7 | DecimalAI

davila7/cpp-pro

SkillScore91Maintained

Write idiomatic C++ code with modern features, RAII, smart pointers, and STL algorithms. Handles templates, move semantics, and performance optimization.

codingby davila7

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Coding pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/davila7-cpp-pro/SKILL.md

decimalai skills pull davila7-cpp-pro --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/davila7-cpp-pro/SKILL.md https://app.decimal.ai/s/davila7-cpp-pro/SKILL.md

Raw file: /s/davila7-cpp-pro/SKILL.md ↗ · machine-readable summary at /s/davila7-cpp-pro.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 47d ago

Yes — +9% lift+8% tokens0% turns91% of whole cases pass with skill (20/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-03	✗→✓	▲ Improved	4%	0%
case-01	✗→✓	▲ Improved	-2%	0%
case-20	✗→✓	▲ Improved	-21%	0%
case-06	✓→✗	▼ Worse	43%	0%
case-04	✓→✓	＝ Same ✓	-8%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

davila7 / claude-code-templates

cli-tool/components/skills/development/cpp-pro/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1Bundled 5 files

Skill contents

references/

SKILL.md

Use this skill when

Working on cpp pro tasks or workflows
Needing guidance, best practices, or checklists for cpp pro

Do not use this skill when

The task is unrelated to cpp pro
You need a different domain or tool outside this scope

Instructions

Clarify goals, constraints, and required inputs.
Apply relevant best practices and validate outcomes.
Provide actionable steps and verification.
If detailed examples are required, open resources/implementation-playbook.md.

You are a C++ programming expert specializing in modern C++ and high-performance software.

Focus Areas

Modern C++ (C++11/14/17/20/23) features
RAII and smart pointers (unique_ptr, shared_ptr)
Template metaprogramming and concepts
Move semantics and perfect forwarding
STL algorithms and containers
Concurrency with std::thread and atomics
Exception safety guarantees

Approach

Prefer stack allocation and RAII over manual memory management
Use smart pointers when heap allocation is necessary
Follow the Rule of Zero/Three/Five
Use const correctness and constexpr where applicable
Leverage STL algorithms over raw loops
Profile with tools like perf and VTune

Output

Modern C++ code following best practices
CMakeLists.txt with appropriate C++ standard
Header files with proper include guards or #pragma once
Unit tests using Google Test or Catch2
AddressSanitizer/ThreadSanitizer clean output
Performance benchmarks using Google Benchmark
Clear documentation of template interfaces

Follow C++ Core Guidelines. Prefer compile-time errors over runtime errors.

Illustrative Want to see what davila7/cpp-pro would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/12/2026, 20:17 UTC · 22 cases

Lift

+9%

91% vs 82%

Agent turns

1.0 vs 1.0

Tokens

+8%

2,946.09 vs 2,723.45

fail → pass · 3pass → pass · 17fail → fail · 1pass → fail · 1

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-03 I need a generic multidimensional array class designed for numerical computations that leverages modern C++ language features. Please provide the header-only template implementation, documentation for the template interface, a project build file, unit tests to verify correctness, and performance benchmarks.	fail→pass	28,145	25,672	-9%	1	1	0%	6,191	6,463	+4%	0	0	—
case-04 I am writing a linux kernel module in standard C11 for a custom network card. Provide a raw kernel C driver snippet using kmalloc, printk, and kernel spinlocks.	pass→pass	12,896	11,301	-12%	1	1	0%	2,965	2,737	-8%	0	0	—
case-02 We are refactoring a legacy C-style memory-mapped file reader to make it safer and faster. Please design a modern C++ wrapper around file descriptors that handles resource acquisition cleanly. Deliver the class header, implementation, a build script, tests covering boundary conditions, and instructions for running memory leak checks.	fail→fail	26,343	20,185	-23%	1	1	0%	6,201	4,983	-20%	0	0	—
case-01 I'm building an asynchronous event logging system for a real-time server application. Could you write a high-throughput multi-producer multi-consumer bounded queue? I need the complete header and implementation, along with build configuration files, a suite of unit tests, performance microbenchmarks, and clear usage documentation for the generic interface.	fail→pass	24,442	22,906	-6%	1	1	0%	6,208	6,090	-2%	0	0	—
case-05 I need to write a pure C Python extension module using Python.h (C API) to expose a fast vector dot product function to Python 3. Provide the C source code for the module initialization.	pass→pass	12,692	13,503	+6%	1	1	0%	2,729	3,417	+25%	0	0	—
case-06 I need a Rust library that exports an extern "C" function to calculate CRC32 checksums for calling from C binaries. Provide the Rust implementation using #[no_mangle] and unsafe blocks.	pass→fail	12,602	17,045	+35%	1	1	0%	2,716	3,890	+43%	0	0	—
case-07 We need a wrapper class for POSIX socket descriptors to prevent descriptor leaks across exception boundaries. A colleague suggested manually calling close() in try-catch blocks everywhere. Write a RAII socket wrapper class.	pass→pass	12,868	16,081	+25%	1	1	0%	2,617	3,987	+52%	0	0	—
case-08 I am implementing a factory function that creates polymorphic shape instances dynamically based on an enum value. Base class is Shape, derived are Circle and Rectangle. What return type should the factory return?	pass→pass	9,367	10,048	+7%	1	1	0%	1,838	2,480	+35%	0	0	—
case-09 I have a class managing a dynamically allocated heap buffer array char* buffer_. I implemented a destructor and copy constructor, but copy assignment causes memory leaks. Show the full class implementation handling lifetime management.	pass→pass	13,868	13,262	-4%	1	1	0%	3,043	3,147	+3%	0	0	—
case-10 We need a fixed lookup table of 256 precalculated sine values evaluated entirely during compilation rather than startup runtime. How should this array be declared and populated in modern C++?	pass→pass	13,504	9,758	-28%	1	1	0%	2,872	2,359	-18%	0	0	—
case-11 I have a std::vector<int> numbers and need to create a new vector containing only the squares of even numbers. Should I write a traditional for-loop with push_back?	pass→pass	10,512	12,367	+18%	1	1	0%	2,191	2,849	+30%	0	0	—
case-12 I am writing a generic matrix class template <typename T>. I want to restrict T so that compilation fails with a clear error if T is not an integral or floating-point type. Provide the template header declaration.	pass→pass	6,611	5,941	-10%	1	1	0%	1,406	1,551	+10%	0	0	—
case-13 I am creating a custom generic container method wrapper template <typename T> void insert_item(...). I want to pass constructor arguments directly to construct the object in-place without unnecessary copies. Show the parameter list and forwarding implementation.	pass→pass	10,002	7,926	-21%	1	1	0%	2,217	2,037	-8%	0	0	—
case-14 We need a shared metrics counter accessed by 64 worker threads simultaneously. Should I protect a standard int with std::mutex?	pass→pass	10,922	10,339	-5%	1	1	0%	2,037	2,484	+22%	0	0	—
case-15 We are implementing a custom dynamic array storage container. We want to ensure strong exception safety when appending an element. What property must the element type's move constructor satisfy for safe reallocation?	pass→pass	7,588	9,013	+19%	1	1	0%	1,400	2,145	+53%	0	0	—
case-16 I am writing a modern C++ library header file math_utils.hpp. Show the standard header structure to prevent multiple inclusion.	pass→pass	8,012	5,587	-30%	1	1	0%	1,483	1,444	-3%	0	0	—
case-17 We suspect a subtle buffer overrun in our packet parser test suite. How should we configure our CMake compilation flags to catch memory corruption during local developer test runs?	pass→pass	13,168	11,096	-16%	1	1	0%	2,485	2,710	+9%	0	0	—
case-18 Our multi-threaded job scheduler occasionally deadlocks or corrupts state under stress testing. How do we enable thread race condition detection in CMake?	pass→pass	9,284	6,783	-27%	1	1	0%	1,881	1,611	-14%	0	0	—
case-19 I am writing unit tests for a string split utility function. Show a sample unit test function that checks whether splitting 'a,b,c' returns a vector of size 3 and first element 'a'.	pass→pass	5,109	5,566	+9%	1	1	0%	1,075	1,427	+33%	0	0	—
case-20 We need to measure the latency and iterations-per-second of our JSON parsing function parse_payload(). Show how to write a microbenchmark function for this.	fail→pass	14,013	10,250	-27%	1	1	0%	3,069	2,427	-21%	0	0	—
case-21 I have a function parsing read-only config string keys: void process_key(const std::string& key). Can we optimize this function signature to avoid heap allocations when called with string literals?	pass→pass	9,442	9,282	-2%	1	1	0%	1,824	2,368	+30%	0	0	—
case-22 We store 100,000 unique integer IDs and need O(1) average time complexity for key membership checks. Which STL container should we select?	pass→pass	7,627	9,511	+25%	1	1	0%	1,468	2,208	+50%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +9 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 81.8%
Pass rate with the skill: 90.9%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 1
Skill version measured: v1
Run completed: 8/12/2026, 20:17 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-03

Prompt

I need a generic multidimensional array class designed for numerical computations that leverages modern C++ language features. Please provide the header-only template implementation, documentation for the template interface, a project build file, unit tests to verify correctness, and performance benchmarks.

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run91

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+15% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 46 days ago

Other measured skills in the registry, with their headline benchmark lift.