jr-explain
jrlprost/claude-code-mods
Every shell command Claude asks permission for comes with a plain-English explanation and a risk level: low, medium or high.
6 mods
jrlprost/claude-code-mods
Every shell command Claude asks permission for comes with a plain-English explanation and a risk level: low, medium or high.
jrlprost/claude-code-mods
Keeps oversized images out of the conversation: an image over 7800 px is read through a resized copy, and browser viewports over the limit are refused.
jrlprost/claude-code-mods
When a reply ends by announcing work ("I'll…", "Let me…", "je vais…") and no tool ran in the turn, Claude is sent back once to actually do it.
jrlprost/claude-code-mods
At the end of a turn, one toast when the code Claude wrote added debug statements (console.log, print, debugger) or TODO/FIXME/HACK lines.
jrlprost/claude-code-mods
Checks every package install against its registry first: names that do not exist (hallucinated or typosquatted) are refused, very new packages are held back or flagged, install scripts are flagged.
jrlprost/claude-code-mods
Replaces API keys, tokens, private keys and database passwords in tool results with placeholders before Claude reads them. Emails and IPs too, if you ask.
No tracking cookies, no third-party scripts. How we handle data