Verified 2026-08-05 What was checked, and when — not a guarantee it is still true.
Writing skills that survive Cowork
Claude Cowork runs your skill in a sandbox with rules that differ from the Claude Code CLI in ways that will silently break an otherwise-correct skill. This site covers the seven failure modes that actually show up: wrong path forms, deliverables that vanish on delivery, delete calls that fail with a permission error, a shell that cannot see your environment, runtime detection that reports the wrong answer, plugin paths that resolve differently per consumer, and sub-agent assumptions that do not hold. Every claim is derived from shipped binaries and live sessions, and is stamped with the build it was checked against. Start with the contract page if you want the short version.
What does my skill's shell see? The shell your skill runs commands in is sealed: environment variables you set on the host do not reach it, and neither do exports written by hooks.
How should my skill detect that it is running in Cowork? A naive environment-variable check false-negatives in exactly the configuration that matters most, because the part of your skill that runs shell commands runs somewhere with none of those markers set.
What can change without any version bump? A large amount of Cowork's behaviour is served from the operator's side rather than shipped in the application, which means it can change with no new version, no release note, and no signal to your skill.