# Foundups Result Interpreter

> Summarize raw test output, logs, or diffs into a compact worker-friendly report for FoundUps handoff. Use after execution upstream returns artifacts.

- Skill: `foundup/foundups-result-interpreter` (Agent Skill)
- Install (CLI): `npx skillmds@latest add foundup/foundups-result-interpreter`
- Raw SKILL.md: https://api.skillmd.com/api/skills/foundup/foundups-result-interpreter/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing
- Author: Foundup (https://skillmd.com/u/foundup)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/foundup/foundups-result-interpreter

---


# 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_title` from packet.

## Steps

1. If input empty → `status` `unknown`, `headline` “No input”.
2. Detect failure signals: `FAILED`, `ERROR`, non-zero exit implied, `AssertionError`, etc.
3. Detect pass signals: `passed`, `OK`, all green summary.
4. `failures`: max 5 strings, each max 120 chars (first errors only).
5. `next_action`: one imperative for **upstream** (e.g. “Re-run tests after fixing import in X”).

## Output format

Emit **only** this JSON:

```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`**

```json
{
  "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`**

```json
{
  "status": "pass",
  "headline": "All tests passed",
  "failures": [],
  "next_action": "Proceed to merge review per WSP 7"
}
```