Show HN: Reprompt – Analyze what you type into AI tools, not what they output (opens in new tab)

(github.com)

3 pointsLuxBennu1mo ago3 comments

3 comments

I ran this on my own prompt history and three things surprised me. found 3 API keys buried in copy-pasted stack traces (`reprompt privacy`). 35% of my agent sessions had error loops -- the agent retrying the same failing approach 3+ times (`reprompt agent`). And 50-70% of my conversation turns were filler like "ok try that" (`reprompt distill`).

    pip install reprompt-cli
    reprompt scan && reprompt

Everything runs locally -- zero network calls, zero telemetry. Also works as an MCP server and GitHub Action.

kiyeonjeon1mo ago

Love the "no LLM calls" approach. Scoring prompts in <1ms locally is exactly the right tradeoff. Most tools overcomplicate this.

LuxBennuOP1mo ago

Thanks! Turns out structural signals get you surprisingly far. An LLM catches more, but speed is the feature.

j / k navigate · click thread line to collapse

3 comments

LuxBennuOP1mo ago

    pip install reprompt-cli
    reprompt scan && reprompt

Everything runs locally -- zero network calls, zero telemetry. Also works as an MCP server and GitHub Action.

kiyeonjeon1mo ago

Love the "no LLM calls" approach. Scoring prompts in <1ms locally is exactly the right tradeoff. Most tools overcomplicate this.

LuxBennuOP1mo ago

Thanks! Turns out structural signals get you surprisingly far. An LLM catches more, but speed is the feature.

j / k navigate · click thread line to collapse