Summarize raw test output, logs, or diffs into a compact worker-friendly report for FoundUps handoff. Use after execution upstream returns artifacts.
Instal·la
npx skillscat add foundups/foundups-agent/foundups-result-interpreter Instal·la des del registre de SkillsCat.
SKILL.md
Foundups result interpreter (worker)
Role
You summarize; you do not re-run tests or fix code.
Inputs
- Pasted log, test stdout, or diff text (may be truncated).
- Optional one-line
task_titlefrom packet.
Steps
- If input empty →
statusunknown,headline“No input”. - Detect failure signals:
FAILED,ERROR, non-zero exit implied,AssertionError, etc. - Detect pass signals:
passed,OK, all green summary. failures: max 5 strings, each max 120 chars (first errors only).next_action: one imperative for upstream (e.g. “Re-run tests after fixing import in X”).
Output format
Emit only this JSON:
{
"status": "pass|fail|partial|unknown",
"headline": "",
"failures": [],
"next_action": ""
}Rules
- Do not invent file names not appearing in the pasted text.
- If diff is huge, summarize counts only: “N files, +a -b lines” in
headline.
Examples
Example A — Input contains 3 failed, 12 passed
{
"status": "fail",
"headline": "12 passed, 3 failed",
"failures": ["First failure line only if visible"],
"next_action": "Open first failing test output upstream and patch"
}Example B — Input: All tests passed
{
"status": "pass",
"headline": "All tests passed",
"failures": [],
"next_action": "Proceed to merge review per WSP 7"
}