AgentWing
Check every AI agent action before it runs
AgentWing launched on May 28, 2026 with 4 votes, #2,419 of 20,538 launches that month and more than 75% of that year's launches.
What it does
AgentWing checks AI-agent actions before they run. Send the proposed action — file read, shell command, package install, code edit, or API call — to /api/v1/check-action. Get back one of five decisions: allow, block, sandbox_required, approval_required, or restore_point_required. Every check produces an auditable receipt. Open-source (Apache-2.0) runtime control layer with BYOK E2B sandbox routing.
AgentWing checks and routes AI-agent execution before tools run: policy decisions, approvals, sandbox routing, restore points, logs, and audit receipts.
AgentWing sits between an AI agent and the actions it wants to take. Before it reads a file, runs a command, installs a package or ships to production, it asks — and gets back one of five answers, plus a receipt. Every action gets exactly one. The decision is deterministic — the same action always produces the same answer, and the answer always names the rule that produced it. Genuinely read-only work is not obstructed. Every segment of a command must be safe, so a safe prefix cannot launder what follows it. Secret reads and irreversibly destructive commands. No custom policy can downgrade these — the floor is not configurable. Deploys, payments, external messages, and anything the engine…from agentwing.gpmai.dev
Does a similar job
all alternatives →More ai this month
the category →



Paintings by AI models: programs run through a simulation of oil paint.
AI · 20d ago · stillwet.art

PixVerse R2▲381A real-time world model you can explore and change
AI · 28d ago · world.pixverse.video
Source: Product Hunt launch ↗
Launched alongside, May 2026
the whole month →

Parallel agents, diff reviewer, and multi-model comparisons
Dev tools · May 2026 · kilo.ai



Hey HN, Henry here from Cactus. We open-sourced Needle, a 26M parameter function-calling (tool use) model. It runs at 6000 tok/s prefill and 1200 tok/s decode on consumer devices. We were always frustrated by the little effort made towards building agentic models that run on budget phones, so we conducted investigations that led to an observation: agentic experiences are built upon tool calling, and massive models are overkill for it. Tool calling is fundamentally retrieval-and-assembly (match query to tool name, extract argument values, emit JSON), not reasoning. Cross-attention…
Life & fun · May 2026 · github.com

🌱 Private, quiet space for thinking. Simple app for .md files. - zakirullin/files.md
Dev tools · May 2026 · github.com