{"version":"2.1.317","anchor":"cli-allow-rules-parsed-with-behavior-skills-entriesanchor","canonical_anchor":"cli-allow-rules-parsed-with-behavior-skills-entriesanchor","heading":"Tools allowed with --allowedTools are now read as allow rules","tier":"notice","area":"Permissions","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.317\/e\/cli-allow-rules-parsed-with-behavior-skills-entriesanchor","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.317","markdown":"### Tools allowed with `--allowedTools` are now read as allow rules\n\nRules passed with `--allowedTools` are now read specifically as allow rules, which may change how some rules are understood\n\n**Unclear.** It is not clear which `--allowedTools` rules, if any, are now read differently.\n\n**What**\n\nThe `--allowedTools` flag, an option you add when starting Claude Code, lists tools that run without asking for permission. Claude Code now reads each rule given there specifically as an allow rule, instead of with the general rule reader it used before.\n\nThis build also has two smaller changes to plugins, which are add-ons that extend Claude Code: skills from plugins load in a slightly different way, and Claude Code records a new step for running plugin hooks.\n\n**Why**\n\nThe new reading may change how some rules passed to `--allowedTools` are understood.\n\n- Area: Permissions\n- Tier: You'll notice\n- Useful: 2\/5\n- Signal: 1\/5\n- Scope: individual\n- Heads-up: no"}