Line of inquiry
Inquiring lines›How do we ensure safety, alignment…›What security vulnerabilities enab…›this line of inquiry
How should agent systems validate and persist generated code artifacts?
A broader line of inquiry — a family of 38 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 38
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- Are durable shared code artifacts better than per-task harness patches?
- How do agents decide which created code deserves long-term persistence?
- How should agents decide which created code is worth persisting?
- How do agents decide which created code should persist versus disappear?
- Can skill libraries prevent redundant narrow artifacts from proliferating?
- Can skill validation through testing prevent unreliable programs from accumulating?
- How should harness infrastructure validate code that agents generate themselves?
- What permission models govern code execution within agent skills?
- How should AI skills be created and managed like software artifacts?
- How should human oversight apply to persistent agent-authored code?
- When should agent-created code be promoted into permanent harness infrastructure?
- How do agent-created code artifacts become part of harness infrastructure?
- Can one-off agent code be safely promoted to durable infrastructure?
- Does held-out validation prevent skill document edits from drifting or accumulating harm?
- What lifecycle management prevents in-loop skill creation from bloating an agent?
- What makes durable code artifacts more valuable than per-task harness patches?
- Can disposable agent-authored code be distinguished from reusable infrastructure?
- How should skills be trusted and installed on sharing platforms?
- What metadata properties make code-derived skills auditable and comparable to their original source?
- What makes persistent, shared code artifacts from agents hard to manage at scale?
- How do artifact families differ in matching verification scope to repair capability?
- Why treat tutorial videos as a separate supply line from agent trajectories?
- How do skills authored in-loop validate faster than offline generated skills?
- What makes skills suitable for retrieval and chaining in repositories?
- How do agents retrieve and compose skills from hierarchical multimodal wikis?
- How does source-blind reconstruction verify that extracted skills are specific enough to be reusable?
- How do composite rewards attribute curation outcomes to specific skill library changes?
- Can agents acquire new skills online when offline skill coverage runs out?
- Does inspectable skill artifacts guarantee the behavior matches the person it claims to ground?
- Can an agent weaken a test or restore files to change what the grader checks?
- Can curator modules trained on one executor transfer to entirely different agent backbones?
- How do capability tracks and behavior tracks stay separable during skill deployment?
- Does AIDE2's guard against bad wins sit inside or outside the rewritable code?
- Did AIDE2's rewrites solve problems on a human checklist or search artifacts?
- Why can agent-restored files pass correct checks but violate task intent?
- Why does embedding research tools in coding assistants improve reliability?
- How do skill libraries from human resources compare to hand-written skill libraries?
- What makes a distilled skill verifiable and ready for agent execution?