Lite effort produces a chat answer, not a report
At Lite effort the research skills deliberately answer in chat with citations and no report document. Switch the Agent control to Basic or Deep and resend to get the report.
The request read as a question, not a task
"What does the literature say about X?" may be answered in chat. Ask for the deliverable: "write a report", "produce a figure", "save the results as a table" — or use the slash command, which always targets an artifact.
The skill ran, but the reviewer rejected the output
Occasionally the reviewer finds the draft unsupported and the turn ends with "no result". Resend the same request; if the data or sources genuinely don't support the deliverable, the answer explains what is missing (for example, a manuscript can't be reviewed if there is no manuscript in the session).
The file has no extension or opens as plain text
Artifacts are named by the agent; the server adds the right extension for the type. If a report shows as raw Markdown, open it from the artifact list (not the code view) — the panel renders .md files.
The artifact list is empty after a resumed turn
Reload the page. Artifacts written before an interruption are in the session even if the answer that lists them was cut short — check the Artifacts page filtered by this session.
Still nothing
Send the request with an explicit skill tag and the deliverable named:
[Use skills: statistics] Test the group difference in PSS-10 change scores and save a statistical report.
If that fails too, get help with the session link.