Rule reference · Output specification

Missing output format

  • missing_output_format
  • info in this example
  • Output specification
  • stage · Output format analysis

Fires when: No output format named.

Format is the part of the output your code or reader depends on most. Left unspecified, it changes between runs: prose one time, a list the next, a table after that.

Name the exact shape: "Return JSON with keys…", "a Markdown table with columns…", "3 bullet points". For machine-read output, also say what to return when data is missing.

Example that triggers it

Analyze the customer reviews below and identify the three most common complaints and what customers say about each of them.

<reviews>
{{reviews}}
</reviews>

info No output format is specified.

Why it matters. Without a format the model chooses one, which makes the output harder to parse or compare across runs.

Fix. Say exactly what to return, e.g. "Return JSON with keys title, summary, tags" or "Answer in 3 bullet points".

Quality 97/100 · 30 tokens · 1 findings

Fixed version

Analyze the customer reviews below and identify the three most common complaints. Return a Markdown table with the columns complaint, number of reviews, and example quote.

<reviews>
{{reviews}}
</reviews>

Quality 100/100 (+3) · 39 tokens · 0 findings

Resolved by the fix

  • missing_output_format: No output format is specified.
Open the example in the workspaceOpen the fixed version

Scores, token counts (GPT-4.1 tokenizer), and findings on this page are computed by the engine when the site is built.

How it affects the score

Each finding subtracts a fixed penalty from 100: critical 30, warning 10, info 3. This rule counts against the Output specification dimension. See the scoring model for the full formula.

All 20 rules →