Skip to main content
Names resolve to archetypes first, then to saved packs. Packs are looked up in ./scenarios and then <data_dir>/scenarios, by their name field or by a path ending in .yaml or .yml.

​
scenario list

Which archetypes bind to this sandbox’s API, with the reason for each that does not, plus saved packs. -v shows which operations each archetype bound to. --format json prints the same payload the MCP scenario_list tool returns: archetypes with applicable and reason, packs and endpoints. Works offline. See Archetypes.

​
scenario check

The text output ends with exercised N of M declared operations (unexercised: …), the spec’s operations the run’s request steps reached. Exit 0 when every scenario passes, 1 when any assertion fails, 2 when a step could not execute or the run could not be set up. No running server is needed; the run builds its own engine. See Running scenarios. pikopod scenario run is the previous spelling. It still works this release, prints the same output, and adds one line on stderr naming scenario check.

​
scenario create

Compile a plain-English description into a pack, grounded against the operations your API actually has. Needs a model key. Prints the draft and asks before saving; --yes saves without asking. See Plain English.

​
scenario from-drift

Pin a drift’s baseline contract as a scenario and replay it immediately. Exit 1 when the replay fails, which confirms the sandbox already reproduces the drifted behaviour. See Generated packs.

​
scenario from-recordings

Generate a pack from a window of recorded traffic: requests in recorded order, status assertions, and producer-to-consumer value chains for identifier-shaped values. Default name traffic-<upstream>.

​
scenario reproduce

Moved to pikopod reproduce. pikopod scenario reproduce still works this release and prints one line on stderr naming the new spelling.