Basic command
lladar report evaluation.json --output report.md--evaluation points to the previous evaluation output. --output names the Markdown report to create.
What the report includes
- What was tested and the overall result.
- Per-question judgments and their reasons.
- How the totals were calculated from the saved results.
- The scope and limits of this run.
Use it to improve the Agent
Do not treat the report as only a score. Group failed questions: is the source material unclear, did the Agent ignore a rule, or did the test miss an important real-world situation? Improve the relevant part, then run the same workflow again.
Use a custom report Skill
A SKILL.md can request your team's sections, tone, or review fields. The individual results and calculated totals remain available so readers can trace a conclusion back to the evidence.
lladar report evaluation.json --skill skills/compliance-summary --output report.md