Ferhat Atagün

    0followers

    Full-stack developer from Turkey. I build hangihesaplama.com, a Turkish calculator platform, and open-source tooling for Claude Code such as rules-bench, which measures what a rules or skill pack costs in context.

    6 skills
    0 uses
    Joined Nov 2026

    Skills by Ferhat Atagün (6)

    Tool Design Sandbox — design Claude tools before building them

    by Ferhat Atagün

    Free

    Review Claude tool schemas and simulate the agent loop with mocked tool results, including errors and bad payloads, before you write any implementation. Reports which tools to keep, rename, reword, split or drop.

    1
    0

    Prompt A/B Test — compare prompts by cost, speed and output

    by Ferhat Atagün

    Free

    Compare two Claude prompts or models on the same inputs using output, cost and latency instead of gut feeling. Gives an experiment recipe, per-input ratios and a verdict with the trade-off spelled out.

    1
    0

    Agent Trace Debugger — replay a Claude agent step by step

    by Ferhat Atagün

    Free

    Turn a Claude tool-use conversation (messages array) into a readable timeline of thoughts, tool calls and results, and flag errors, orphaned calls, loops and unrecovered failures with a concrete fix for each.

    1
    0

    Claude API X Ray — cache, tokens and cost per call

    by Ferhat Atagün

    Free

    Split any Claude API response into cache read, cache write, uncached input and output, price each segment, and find out why prompt caching is not hitting. Works from a usage block; includes a dated model price table.

    1
    0

    Prompt Preflight — tokens, window and cost before you ship

    by Ferhat Atagün

    Free

    Check a Claude prompt before it ships: estimated or exact input tokens, position in the context window, cost per call and per 1,000 calls, and where prompt caching would (and would not) hit. Includes a dated model and price table.

    1
    0

    Context Cost Audit — what does your Claude Code pack cost?

    by Ferhat Atagün

    Free

    Measures what a CLAUDE.md, rules folder or skill/agent pack costs in context before your first prompt (startup) versus only when triggered (on-demand), then tells you what to trim.

    1
    0