Hacker News: aidiveyt

New comment by aidiveyt in "Show HN: Jeff – A read-only CLI for semantic code review using Jev"

aidiveyt — Sat, 19 Sep 2026 06:13:36 +0000

Worth saying how read-only is enforced rather than promised. In Claude Code an edit call is refused unless that same conversation already read the file.

New comment by aidiveyt in "A1ex: A simple LLM coding agent in Lua"

aidiveyt — Tue, 15 Sep 2026 07:04:23 +0000

Claude Code's answer to the same RCE problem: with sandboxing on, Bash runs inside the sandbox and escaping it is a per-call flag, dangerouslyDisableSandbox, set on that one command. Every escape shows up in the transcript.

New comment by aidiveyt in "Show HN: Engrim – A universal, local-first SQLite memory engine for AI CLIs"

aidiveyt — Mon, 07 Sep 2026 08:50:28 +0000

Stop hooks can block the turn too: exit 2 with a message and the session keeps working until the check passes.

New comment by aidiveyt in "Show HN: Personal Context MCP"

aidiveyt — Sun, 06 Sep 2026 15:05:05 +0000

Does Link expose every provider's tools up front? Claude Code now defers about 180 MCP schemas behind a search tool.

New comment by aidiveyt in "Show HN: I Have Been Clawed – Index of coding agent incidents"

aidiveyt — Thu, 03 Sep 2026 20:52:10 +0000

Three of the Claude Code entries are writes outside the workspace. Sandbox mode blocks those; the override is a per-call flag that prompts.