AI / OpenClaw Interview Questions
What is the difference between a skill's self-declared permissions and independently verified sandbox enforcement?
| Self-declared permissions | Independently verified enforcement |
| Listed in the skill's SKILL.md manifest by its author | Enforced at runtime regardless of what the manifest claims |
| Provides capability transparency, what the skill says it needs | Provides an actual guarantee about what the skill can do |
| Trusts the skill author to declare accurately and completely | Doesn't depend on trusting the author's declaration at all |
| Not currently independently checked against real behavior | Requires sandboxing or a runtime policy engine to enforce |
The gap between these two is a documented governance concern: OpenClaw currently has no independent runtime verification confirming that a skill's actual behavior matches its declared permissions, meaning a malicious skill could simply declare narrow permissions while behaving differently once running with operator-level trust.
What does self-declared permission provide?
Is a skill's real behavior currently checked against its declared permissions?
More Related questions...
