- Home
- Skills
- Agents & Orchestration
- bend-skills
Works with the AI Agents you already use
Bend Skills
Develop, benchmark, and verify massively parallel functional programs using Bend Classic and Bend 2.
Free
bend-skills
Example session with this skill installed
Write a Bend classic program to calculate Fibonacci(20) using a recursive divide-and-conquer approach, then benchmark it in both rs and c modes.
- Read your context and instructions
- Compiled the bend-
I have created the recursive Fibonacci program and performed a benchmark across sequential (rs) and parallel (c) modes.
Recursive Fibonacci Benchmark
# fib_recursive.bend
def fib(n):
if n == 0:
return 0
elif n == 1:
return 1
else:
# Divide-and-conquer: HVM reduces these branches in parallel
return fib(n - 1) + fib(n - 2)
def main():
return fib(20)
| Metric | Rust (Sequential) | C (Parallel) |
|---|---|---|
| Median Time | 0.045s | 0.012s |
| Min Time | 0.042s | 0.010s |
| Result | 6765 | 6765 |
Next steps
- Run
bend_run(name="fib_recursive", mode="cu")if you have an NVIDIA GPU to test massively parallel reduction. - Increase the input value to
fib(32)to see a more significant speedup as the interaction graph grows. - Use
bend_read_program(name="fib_recursive")to verify the source before deployment.
Connects securely to your tools. The creator never sees your data.
About this skill
Writing massively parallel code often results in a trade-off between performance and developer velocity. This skill enables the creation, execution, and benchmarking of functional programs in Bend, a language designed for implicit parallelism across CPUs and GPUs via interaction combinators.
What it does
- Multi-version support provides dedicated workflows for both Bend classic (0.2.x) and Bend 2 (2.x) syntax.
- MCP integration uses
bend-mcpto write, compile, and run programs directly within your environment. - Hardware benchmarking compares execution across Rust/CPU, C/Parallel, and CUDA/GPU modes.
- Verification and proofing validates syntax for classic programs and executes mathematical proofs for Bend 2 logic.
- Performance analysis identifies code patterns like divide-and-conquer that benefit from implicit parallelism.
How it works
bend_check_installconfirms the installed version (Classic vs v2) and hardware capabilities like CUDA availability.bend_write_programsaves source files to the workspace for execution or compilation.bend_checkperforms static analysis to catch syntax errors or proof failures before runtime.bend_runexecutes the code using the appropriate backend mode (rs, c, or cu).
Frameworks & tools
Supports bend-lang (0.2.x), hvm (2.0.x), and the new bend (2.x) compiler. Works with CUDA for NVIDIA GPU acceleration and Node.js/C for transpilation targets.
Why this beats prompting it yourself
Generic models frequently hallucinate and mix syntax between the incompatible 0.2.x and 2.x versions. This skill enforces version-specific rules, such as affine type constraints in Bend 2 and loop-specific fork usage in Classic, ensuring code actually compiles.
Use cases
- Porting recursive algorithms to massively parallel GPU kernels.
- Verifying mathematical properties of functional programs using Bend 2 proofs.
- Benchmarking speedups between sequential Rust and parallel C backends.
- Implementing IO-driven CLI tools using Bend's monadic effects.
Known limitations
Bend 2 requires explicit type annotations and lacks a native if keyword. CUDA execution requires nvcc and compatible hardware; otherwise, the skill falls back to CPU execution.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Bend 2 requires explicit type annotations and lacks 'if' (uses match). GPU modes require nvcc and NVIDIA hardware. No cross-mixing of 0.2.x and 2.x syntax.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 27 days ago
- Free with an Agensi account
- Accounts
Anthropic
GitHub
Llvm
Bend Lang
- Software
- Bend MCP
Permissions
Allowed Hosts
Skill/Conector IA