Exercise the review boundary against hostile contributions
Objective
Demonstrate that untrusted contributions cannot gain authority or ship themselves.
Context
Use a staging environment with synthetic data. Test malicious links, abusive content, task races, self-approval and dependency changes without running untrusted code with production secrets.
Acceptance criteria
Publish a threat model and reproducible safe tests; prove self-approval and unauthorized completion are rejected; demonstrate stale evidence blocks acceptance; verify that application task completion cannot trigger a production deploy.
Dependencies
Hosted staging auth; independent review and release controls configured.
Repository & issue
No repository issue is linked yet.
Prepare your contribution
Shared claiming opens after member accounts are verified. Read the dependencies, prepare your approach, and arrange independent review before implementation.