The terminal showed the real output. The agent's tool result didn't.
What the report says
A user running Cline 4.0.12 inside a VS Code Remote-SSH session to a Linux VM filed cline/cline#12723 after execute_command silently dropped the real stdout, stderr, and exit status of shell commands it ran, always reporting a generic success back to the model instead.
In practice, this meant a failing Ruff lint check got recorded as passing. In the reporter's own words, the defect "caused a Ruff failure to be treated as passing and inaccurate verification evidence to be committed to the project issue tracker."
The reporter's minimal reproduction was small: a bash one-liner printing distinct stdout and stderr markers and exiting with status 7. The VS Code terminal itself correctly displayed both streams and the exit code. Cline's tool result reported success regardless. The same false-positive behavior reproduced on both 4.0.11 and 4.0.12 — meaning the defect had already survived one release cycle before anyone reported it.
What it doesn't establish
This is one filed, verified, reproducible report, scoped specifically to VS Code's Remote-SSH terminal mode — not a demonstrated failure across Cline's other terminal integrations, and not (yet) a confirmed root cause or fix from the maintainers.
Why it's scored 3.8/low, not higher
No data was lost and nothing was deployed on the strength of the false result. What makes it worth documenting anyway: the core guarantee an agent's self-reported command output should track reality broke silently, in exactly the kind of check (lint, test, build) a developer would otherwise trust without re-running by hand.
Top comments (0)