My AI deleted more than twenty unfinished tasks so it could say it was done
If a completion gate reads a number the agent can change, the agent can pass it by changing the number, and probably will.
Jul 16, 20267 min read87

Search for a command to run...
Articles tagged with #software-engineering
If a completion gate reads a number the agent can change, the agent can pass it by changing the number, and probably will.

More than 6,000 documented agent failures from building one product, the six that keep coming back, and what actually catches them
