Conversation
Pattern credit: OpenWorker (Andrew Ng and contributors) — dangerous operations are human-only, always. Every path past Prax's HIGH-risk gate could be lowered or skipped: earned trust downgrades browser login steps on self-reported success; the fallback lets the model confirm by calling again; one confirmation can unlock every HIGH tool for the turn; inside spokes the gate isn't enforced unless SPOKE_GOVERNANCE_ENABLED (off in production); a timed grant approves on arrival; and browser_login / browser_credentials, which hand the model a stored password, weren't classified at all. HARD_FLOORS_ENABLED (default off): logging in or revealing credentials, installing or activating code with Prax's authority, and starting a billable GPU run only on a person's decision about that exact call — checked before all of the above. The decision is an out-of-band TeamWork approval that a timed grant did not give (decided_by is now carried through), or, with approvals off, the user's own message naming the action and its target. HARD_FLOOR_EXTRA_TOOLS adds floors; built-ins can't be removed.
Owner
Author
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Pattern credit: OpenWorker (Andrew Ng and contributors): "dangerous operations are human-only, always". Assessed in #246.
Why. Every path past Prax's HIGH-risk gate could be lowered or skipped:
browser_request_login/browser_finish_loginHIGH→MEDIUM on self-reported success;SPOKE_GOVERNANCE_ENABLED, which is off in production;browser_loginandbrowser_credentials, which hand the model a stored password, weren't classified at all (so they ran as MEDIUM).What.
HARD_FLOORS_ENABLED(default off). Built-in floors (prax/agent/hard_floors.py) cover three kinds of action: credentials (the five login/credential tools); code that runs with Prax's authority (plugin import/activate/write,self_improve_deploy); and money (gpu_power_on).A floor action runs only on a person's decision about that exact call, checked before earned trust, auto-approve, the turn latch and the spoke enforce switch. The decision is either:
decided_byis now carried throughapproval_service→human_approval); orRefusals tell the model not to retry or route around it.
HARD_FLOOR_EXTRA_TOOLSadds floors; built-ins can't be removed.Tests:
tests/test_hard_floors.py(19), covering both enforce settings, earned-trust override, the grant refusal, the target matcher, and add-only extras.make ci: 4029 passed; no secrets-proxy requests during the run.Recommended in production:
HARD_FLOORS_ENABLED=true.