Show HN: AI codebase analyser and auto-fixer
getdebug CLI 0.4.0 ships Python AI-app regex prefilters alongside the JS/TS ones that landed in 0.3.0. Five categories: prompt-injection, unsafe-role-merge, pii-in-prompt, unbounded-stream, unsafe-tool-output. The interesting part is the benchmark we ran while developing them, against the three other tools anyone would compare us to.The four tools, the honest mapBandit (PyCQA) is the Python-OSS standard security linter.
- ▪getdebug CLI 0.4.0 ships Python AI-app regex prefilters alongside the JS/TS ones that landed in 0.3.0.
- ▪Five categories: prompt-injection, unsafe-role-merge, pii-in-prompt, unbounded-stream, unsafe-tool-output.
- ▪The interesting part is the benchmark we ran while developing them, against the three other tools anyone would compare us to.The four tools, the honest mapBandit (PyCQA) is the Python-OSS standard security linter.
Hacker News (AI / LLM) files mainly under ai. We currently carry 2,240 of its stories.
Opening excerpt (first ~120 words) tap to expand
getdebug CLI 0.4.0 ships Python AI-app regex prefilters alongside the JS/TS ones that landed in 0.3.0. Five categories: prompt-injection, unsafe-role-merge, pii-in-prompt, unbounded-stream, unsafe-tool-output. All deterministic. No LLM call. The default getdebug analyze . on a Python repo runs in milliseconds, costs zero, and now covers Python AI-app anti-patterns alongside the secrets + dependency-CVE passes it always ran.The interesting part isn't the patterns themselves — they're straightforward regex on familiar SDK idioms (messages=[{"role": "system"...}], stream=True, subprocess.run(tool_call.input.command)).
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Hacker News (AI / LLM).
Discussion
0 commentsMore from Hacker News (AI / LLM)
