Interrogate

---
name: interrogate
description: "Use for \"interrogate\", \"adversarial review\", \"multi-model review\", \"challenge this\", \"stress test this code\", \"find blind spots\", or \"tear this apart\". Multiple LLM reviewers challenge changes from independent angles."
---

# Interrogate

Read [the pstack-t3 runtime](../pstack-runtime/SKILL.md) before spawning workers, choosing models, scheduling, or isolating work. It maps those steps onto T3's orchestrator tools.

Spawn one reviewer per configured model to adversarially review code changes. Each model gets the same prompt and rubric. The adversarial signal comes from model diversity, not assigned personas.

The deliverable is a synthesized verdict. Do NOT auto-apply changes.

## Step 1, Determine Scope

Identify what to review from context:

- If the user points at specific files or a diff, use that
- If on a feature branch, run `git diff main...HEAD` (or the appropriate base branch) for the full changeset
- If the user's message references recent work, gather the relevant files

Package the diff (or file contents) plus any surrounding context files the reviewers need to understand the code.

## Step 2, State the Intent

Before spawning reviewers, state the intent explicitly. Derive this from:

- The user's message
- Commit messages
- PR description if one exists
- The code itself

Write one clear paragraph. If you're unsure about the intent, ask the user before proceeding.

## Step 3, Spawn Reviewers

Call `orchestrator_capabilities`. Paste that tool result into this quoted heredoc. If the catalog result is large, save it to a temporary file with the host's file tool and pass that path to `--catalog`.

```bash
python3 <pstack-runtime>/scripts/roles.py show --cwd "$PWD" --catalog - --parent "<inheritedProviderInstanceId>/<inheritedModel>" --role "interrogate reviewers" <<'JSON'
<the orchestrator_capabilities JSON>
JSON
```

The quoted heredoc sends the JSON unchanged. The command does not write the catalog into the repository, and parallel children do not share a file. It resolves the `interrogate reviewers` role per [the runtime's Roles section](../pstack-runtime/SKILL.md#roles). One reviewer per seat, labeled Reviewer A, B, and onward. The seat count is the panel size.

Launch all reviewers in a single message, one `delegate_task` call per seat:
- `mode`: `"async"`
- `role`: `"review"`
- `target`: the seat's resolved target. For an `inherit` seat, omit `target` so that reviewer runs on the parent model.
- `clientRequestId`: stable per seat, such as `interrogate-<slug>-a`
- `task`: the filled template below, which is a read-only brief

Retain every returned `taskId`. If `delegate_task` rejects a target, apply the runtime's fallback, spawn on the fallback seat, and say which reviewer changed and why. Do not block the review on a target issue. Never silently drop a seat.

Read `references/reviewer-prompt.md` and fill in the template with:
1. The stated intent
2. The diff or file contents
3. The review rubric from `references/rubric.md`
4. The code-quality lens from `references/code-quality-review.md`

The same filled template goes to all reviewers, so every model applies the code-quality lens.

End the turn and let each completion notification wake you, or call `task_status` with a retained `taskId`. If a reviewer fails or returns nothing usable, proceed with N-1 and record the dropout.

## Step 4, Synthesize

As results come back, build a unified picture:

1. **Parse all findings** from the reviewers
2. **Identify consensus**. Findings raised by 2+ models independently are highest signal.
3. **Identify lone-model findings**. Still worth reading, but weight accordingly.
4. **Deduplicate**. Different models may describe the same issue differently. Merge these and note which models raised it.
5. **Note disagreements**. If one model flags something and another explicitly says the opposite, that's useful context for the verdict.

## Step 5, Lead Judgment

You are the lead reviewer, a pragmatic senior engineer, not a neutral aggregator.

Read `references/lead-judgment.md` for the full framework.

Categorize every finding using these buckets:

- **Act on**. Real issues affecting correctness, security, or maintainability given the actual goals. These would block a real PR.
- **Consider**. Legitimate points, but you're not sure they outweigh the cost of addressing them right now. Worth the user's attention.
- **Noted**. Technically valid but not actionable. Context-dependent, premature optimization, or low-impact given the current stage.
- **Dismissed**. Wrong, nitpicky, or missing context. Brief explanation why.

For each finding, include:
- Which model(s) raised it
- The category (act on / consider / noted / dismissed)
- A one-line rationale for the categorization

## Output Format

Present the verdict in this structure:

### Intent
> [The stated intent paragraph from Step 2]

### Reviewers
- Reviewer [label]: [provider/model, or inherit and the parent model], [N findings] (one bullet per reviewer)
- Fallbacks and dropouts, and whether the models actually differed

### Act On
[Findings that should be addressed. For each: description, which models raised it, why it matters.]

### Consider
[Findings worth thinking about. For each: description, which models raised it, tradeoff involved.]

### Noted
[Valid but low-priority. Brief list.]

### Dismissed
[Rejected findings with brief rationale.]

### Agreement Map
[Where did models agree, where did they diverge, and what does the pattern of agreement/disagreement tell us?]