Updated daily · last run Nov 20, 07:02 UTC
Trending in the bucket
Every AI dev tool the internet posted about, ranked by who's talking about it. Sorted. Mostly.
- In the bucket
- 3.1K
- Posts read
- 42.7K
- Tossed out
- 33.7K
43 tools, ranked by sources, chatter and star growth. How ranking works
- 01
#1
SSH agent proxy that lets remote coding agents run commands on your laptop with policy rules, Touch ID approval, and secret injection.
HN#claude-code#codex#macos1 post · 4 days ago★ 28
Stars:28
- 02
#2
Sandboxed runtime for AI agent fleets with kernel-level policy enforcement, credential brokering, and formally verified policy changes.
GHHN#any-agent#linux#macos2 posts · 28 days ago★ 21.8K +221/wk
Stars:21.8K
+221 this week
- 03
#3
macOS notch app that lets agents request single Bitwarden credentials, approved per request with Touch ID.
GH#macos#any-agent#menu-bar1 post · 3 days ago★ 24
Stars:24
- 04
#4
Finds API keys leaked by AI agents via planted canary keys; includes a Claude Code status line and redaction hook.
HN#claude-code#cli#github-action1 post · 10 days ago★ 19
Stars:19
- 05#5
MCP server giving coding agents a sandboxed, readable scripting runtime with security policy, secrets, taint tracking, audit logs and undo.
HN#mcp-server#any-agent#macos1 post · 10 days ago
no public repo
- 06
#6
Claude Code hooks that block parallel sessions from killing each other's processes, ports, RAM and git branches.
GHHN#claude-code#plugin#macos2 posts · 12 days ago★ 35 +1/wk
Stars:35
+1 this week
- 07
#7
Sandboxes AI coding agents in a macOS VM, showing what ran and reviewing file changes before they reach your project.
GHHN#claude-code#codex#macos3 posts · 28 days ago★ 288 +72/wk
Stars:288
+72 this week
- 08
#8
Zero-setup, zero-latency sandbox for running AI coding agents with least-privilege filesystem, network, and credential policies.
HN#claude-code#codex#opencode3 posts · 23 days ago★ 5.1K +86/wk
Stars:5.1K
+86 this week
- 09
#9
OS-level sandboxing tool from Anthropic that enforces filesystem and network restrictions on processes, built for safer AI coding agents.
HN#claude-code#macos#linux1 post · 27 days ago★ 7.2K +77/wk
Stars:7.2K
+77 this week
- 10
#10
Local privacy proxy for AI coding agents that replaces PII and secrets with pseudonyms before requests leave the machine.
GH#claude-code#codex#any-agent1 post · Oct 21★ 246 +46/wk
Stars:246
+46 this week
- 11
#11
Rootless container runtime and sandbox with no daemon that runs agent tool calls and LLM-generated code in kernel-enforced containers, with
GHHN#mcp-server#claude-code#cursor3 posts · Oct 16★ 493 +16/wk
Stars:493
+16 this week
- 12
#12
Open-source AI agent security platform: inventories coding agents, detects risky actions, and enforces guardrails at agent hooks.
GH#claude-code#codex#cursor1 post · 15 days ago★ 12 +1/wk
Stars:12
+1 this week
- 13
#13
Endpoint detection and response for AI coding agents: local monitoring, optional pre-action blocking, and forensic reconstruction of agent a
GHHN#claude-code#codex#macos3 posts · Aug 12★ 1.2K +10/wk
Stars:1.2K
+10 this week
- 14#14
Runs coding agents in sandboxed Docker containers, intercepting network calls and logging LLM traffic, tool calls, and file changes.
HN#any-agent#docker#macos1 post · Oct 2
no public repo
- 15
#15
CLI that wraps AI coding agent runs with file snapshots, reversible PostgreSQL transactions, and review of outbound effects.
GH#any-agent#macos#linux1 post · Jul 26★ 16
Stars:16
- 16
#16
Local, offline tool that records AI coding agents' file changes, reverts any of them, and alerts on secret reads and dangerous commands.
GH#claude-code#codex#cursor1 post · Jul 26★ 11
Stars:11
- 17#17
HTTP proxy that gates coding agents' outbound requests with allow lists, an optional LLM advisor, and operator approval.
HN#claude-code#any-agent#macos1 post · Jul 14
- 18
#18
CLI tool that hooks into Claude Code to snapshot files before destructive agent shell commands and journal them for undo/redo.
HN#claude-code#cli#macos1 post · Aug 20★ 33
Stars:33
- 19
#19
Kit of Claude Code hooks, settings, status line and audit script that blocks destructive commands, protects .env files and tunes output styl
GH#claude-code#macos#linux1 post · Aug 24★ 37
Stars:37
- 20
#20
A skill and hook bundle for Claude Code that audits dependencies for CVEs, typosquats, and abandoned packages across seven ecosystems.
HN#claude-code#plugin#macos1 post · Aug 25★ 40
Stars:40
- 21
#21
Local MCP gateway with signed policies, one-time approvals and an evidence trail controlling which tool calls AI agents can run.
GH#mcp-server#any-agent#cli1 post · Sep 24★ 13
Stars:13
- 22
#22
A secure runtime for autonomous AI agents that enforces a constitution-derived policy on every tool call.
HN#claude-code#codex#docker1 post · Jul 14★ 612
Stars:612
- 23
#23
A Rust gate between AI coding agents and the shell that previews destructive commands, backs up before running, and enforces a readable poli
HN#claude-code#codex#cursor3 posts · Sep 29★ 27
Stars:27
- 24#24
Mac background app that detects secrets and PII in AI tool prompts and clipboard before they are sent, with local-only detection.
HN#claude-code#cursor#copilot1 post · Jul 16
no public repo
- 25
#25
Local CLI that runs checks and binds signed results to a git-state fingerprint, flagging STALE when code changes after an agent's test run.
GHHN#any-agent#cli#github-action2 posts · Aug 16★ 15
Stars:15