Claude Managed Agents Explained: How the Hosted Runtime Fits Your SDLC
blaizedsouza · x · 2026-09-25
The New Stack breaks down Anthropic's Claude Managed Agents (public beta on Claude Platform since April 8, 2026):
- What it is: a managed agent harness where Anthropic runs the infrastructure — tool-call decisions, isolated code execution, credential handling, and hours-long task tracking — so teams only define what the agent does and what it can reach.
- Single-agent strength: jobs can run for hours, survive dropped connections, and safely touch real systems unattended.
- Multi-agent limit: orchestrating several agents still needs a layer above them to chain steps, keep a human in the loop, and track the whole run — the "AI software factory" platform teams build on top.
Aimed at platform engineering teams wiring agents into the SDLC.
More from coding & agent
- LangChain Launches Managed Deep Agents With Built-in Parallel-Powered Web Search — LangChain · 2026-09-25
- Google's Antigravity SDK now runs open models like Gemma 4 fully offline on local GPUs — HankYeomans · 2026-09-25
- NeurIPS Spotlight Paper DCI: Ditching Top-k Retrieval for grep-Style Agentic Search — stanfordnlp · 2026-09-25
- Opus 5.5 writes full stage lighting rig: real fixture specs, outputs Artnet and sACN — natesiggard · 2026-09-25
- Agent side-loads custom app on DJI RC2, streams drone video to RF-DETR detections in real time — burny_tech · 2026-09-25
- Developer uses Opus 5.5 to stitch together an animated historical timeline — DavidmComfort · 2026-09-25