vlinx-io / vask
Ask a focused question about another vlx-term session without importing its full conversation into the current context. This is the skill form of `vrefer SESSION --ask "QUESTION"`. Use it when the target session is known and the user needs a conclusion, decision, result, reason, or status from it. Read-only: nothing is sent to the target session and nobody is interrupted. Available only inside vlx-term-hosted local sessions.
vask
Use vask as the focused-question form of vrefer --ask. There is no separate shell command: invoke
the existing read-only command with the session reference and the user's question.
vrefer <session> --ask "<question>"If the target is not known, search first:
vsearch <distinctive words...>
vrefer <id-from-search> --ask "<question>"<session> accepts a full id, an id prefix of at least eight characters, an exact session name, or a
unique substring. Quote names containing spaces. If several sessions match, select an id from the
candidate list and retry rather than guessing.
By default the answering agent receives the complete transcript. If the user enabled Summarize first
in Settings → Agents, VelaTerm compresses the transcript with the one globally selected Agent, model,
and reasoning effort, adds relevant original full-text search excerpts from the same session, and then
asks the answering agent. This setting is automatic; the skill command and --ask syntax do not change.
Use --with <kind> only when the user explicitly wants a particular answering agent. It does not
change the configured pre-summary Agent. Use --timeout <seconds> when the default 120-second overall
limit is unsuitable.
The returned text is an agent's reading of another session, not a verbatim transcript. When exact wording
or broader context matters, follow up with vrefer <session> or vrefer <session> --range A:B.
If answering fails, vrefer explains the reason on stderr and returns the transcript as its fallback.
Check stderr before treating stdout as a completed answer.