Docs / Agents: CLI, MCP, skill
Command reference
Every command the a9script tool answers to, from the tool's own tables. A command marked as crossing MCP is offered to an assistant as a tool of the same name; the others stay on the terminal by decision, and the reason is stated.
| Command | What it does | Words | Flags | Over MCP |
|---|---|---|---|---|
a9script login | store a PAT | --server--token | no — it stores a token in a profile FILE, and an MCP connection carries its token in the header it arrived on | |
a9script whoami | who the stored credential is, and where it is bound | yes | ||
a9script setup | turn a first-run setup token into the first administrator | --email--name--password--server--token | no — it turns a setup token into the first account — the one state a door behind a credential cannot be in | |
a9script filestore | read and create the environment's file stores | <action><name> | --config | yes |
a9script files | move files in and out of a file store | <action><store><path> | --from--out--type | no — it moves bytes between the caller's own disk and the store, and a remote caller's disk is not the server's |
a9script entity | read and write the environment's entities, one at a time | <action><kind><name> | --data--file | yes |
a9script apply | upsert a .json file, or a folder of them, by name | <path> | --atomic--dry-run | yes |
a9script describe | what a door takes: its URL, the credential it asks for, the networks it answers, what runs behind it and the shapes it declares | <endpoint> | yes | |
a9script lock | read the environment's edit lock, or give it back | <action> | yes | |
a9script run | follows the run through any pause and exits on the RUN's outcome | <script> | --debug--input--no-wait--step--wait | yes |
a9script probe | fire ONE configured call and read the whole exchange | <function> | --acknowledge-write--input | yes |
a9script sessions | the environment's paused stepped runs | yes | ||
a9script position | where a stepped run stands: the parked call, the statement, the variables in scope | <run> | yes | |
a9script continue | one more external call | <run> | --to-end | yes |
a9script abort | stop a paused stepped run; nothing further is sent | <run> | yes | |
a9script history | a debug run's calls: names, status, verdict | <run> | yes | |
a9script inspect | one call whole: request, attempts, bodies | <run><seq> | yes | |
a9script authoring-pack | how to write a script, for humans and agents | --part | yes |