01 / DECIDE
Architecture decision
Use them when repeatable methods or parallel evidence work reduce context and coordination cost.
Skills carry reusable instructions and resources; plugins bundle extensions; hooks observe lifecycle; subagents own independent tasks. Map input, output, state, tools, and side effects as one observable path, then name the owner of package methods as skills and isolate work with subagents.
02 / OPERATE
Three-step field procedure
- 01
Lock the contract
Write down the caller, data classification, success condition, timeout, cancellation, and ownership. Use them when repeatable methods or parallel evidence work reduce context and coordination cost.
- 02
Run one bounded path
Build one end-to-end path with correlation, typed state, and reversible failure handling. Skills carry reusable instructions and resources; plugins bundle extensions; hooks observe lifecycle; subagents own independent tasks.
- 03
Verify the signal
Turn acceptance into a repeatable fixture, contract test, or browser test. Trigger, extension validation, hook failure, and ownership tests prove every layer behaves within scope.
03 / BOUND
Operational boundary
Audit installed code, trust project hooks explicitly, and never assign conflicting writes to parallel agents.
04 / PROVE
Acceptance evidence
Trigger, extension validation, hook failure, and ownership tests prove every layer behaves within scope.
SOURCE / HTTP
Reproducible source probe
curl -fsSI 'https://docs.x.ai/build/features/skills-plugins-marketplaces' | sed -n '1,5p'