interrogate

An agent skill by cursor, from cursor/plugins. Tags: analysis, code-review, debugging, multi-agent, review.

What it does

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.

Install

With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:

npx skills add cursor/plugins --skill interrogate

Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):

git clone --depth 1 https://github.com/cursor/plugins
cp -r plugins/pstack/skills/interrogate ~/.claude/skills/interrogate

Safety box score

Grade A, minimal risk: 7 of 100 risk points, rated 2026-10-20. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.

This skill orchestrates an adversarial code review: it reads the target repo, may run `git diff`, and spawns model-based reviewers, but it does not modify code or contact external services on its own. The main exposure is limited read access outside the project to `~/.cursor/rules/pstack-models.mdc`, plus the usual broad repository inspection needed to review a change set.

Category levels, from 0 (none) to 3 (high)
Category	Level	Why
Code & shell execution	1 (low)	It can trigger tool actions like `git diff main...HEAD` and Task-based reviewer calls, but there are no shipped scripts, installers, or arbitrary code-eval paths.
Network access	0 (none)	The skill files do not download from or send data to remote hosts; there is no `curl`, API call, or other network sink in the shipped content.
Filesystem reach	1 (low)	It explicitly reads outside the project at `~/.cursor/rules/pstack-models.mdc`, though the access is read-only and otherwise stays within the repo diff/context.
Secrets & credentials	0 (none)	There is no instruction to read credential stores, `.env` files, keychains, browser sessions, or to ask the user for secrets.
Privilege & persistence	0 (none)	No sudo, global installs, service registration, shell startup edits, or OS setting changes are requested.
Instruction hijack surface	0 (none)	The skill contains review orchestration instructions only; it does not tell the agent to ignore safeguards, hide behavior, or follow fetched instructions.
Irreversible actions	0 (none)	It is explicitly non-mutating: `Do NOT auto-apply changes.` The output is a synthesized review, not an action that changes state.
Transparency	0 (none)	All shipped files are plain markdown text; there are no binaries, minified blobs, or runtime-decoded payloads.

Before installing

Check whether `~/.cursor/rules/pstack-models.mdc` contains anything sensitive or machine-specific before letting the skill read it.
Confirm you want the agent to review whole-change diffs via `git diff main...HEAD`, not just the files you named.
Open `SKILL.md` and verify you are comfortable with the skill’s external dependency on local Cursor rules for model selection.

Rated 2026-10-20 by an automated review of 5 files from the skill's repository.

Source

Repository: cursor/plugins (all skills from this repository)
Path: pstack/skills/interrogate/SKILL.md
Branch: main
Collection: cursor
Updated: 2026-09-23

thermos — Launch both thermo-nuclear review subagents in parallel, then synthesize their findings.
no-comments — Spawn Comment Sicko, fix accepted findings, and offer encodings for claimed constraints.
reflect — Spawn three parallel review subagents over the active transcript, surface learnings, and route each to a concrete edit on an existing skill.
thermo-nuclear-code-quality-review — Run an extremely strict maintainability review for abstraction quality, giant files, and spaghetti-condition growth.
thermo-nuclear-review — Comprehensive security and correctness audit of a branch's changes.
why — Use for 'why does X work this way', 'why we picked Y', design rationale, regressions, postmortems, or data-backed thresholds.