You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One thing I would turn on immediately if it existed: a strict operator mode that optimizes for not making the workspace worse.
The contract would be simple:
no crossing repo boundaries unless explicitly authorized
no scratch docs, temp notes, or workflow noise unless they were requested
no pretending partial work is complete
no repeated retries when the evidence says the current path is wrong
no public-facing text until it passes a contamination and polish check
A lot of the real pain is not one bad answer. It is the cleanup tax after drift:
reverted context
confused source of truth
messy public output
wasted tokens on repair instead of progress
I keep thinking the best version of coding-agent tooling is not just smarter generation. It is tighter admission control around what the agent is allowed to touch, keep, retry, and publish.
Curious what other operators would put in that hard-mode contract.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
One thing I would turn on immediately if it existed: a strict operator mode that optimizes for not making the workspace worse.
The contract would be simple:
A lot of the real pain is not one bad answer. It is the cleanup tax after drift:
I keep thinking the best version of coding-agent tooling is not just smarter generation. It is tighter admission control around what the agent is allowed to touch, keep, retry, and publish.
Curious what other operators would put in that hard-mode contract.
Beta Was this translation helpful? Give feedback.
All reactions