why
An agent skill by cursor, from cursor/plugins. Tags: analysis, debugging, research, review, strategy.
What it does
Use for 'why does X work this way', 'why we picked Y', design rationale, regressions, postmortems, or data-backed thresholds. Discovers available MCPs and queries each evidence category (source control, issue tracker, long-form docs, real-time chat, infrastructure observability, error tracking, product analytics warehouse) in parallel, then returns a cited read on decisions and tradeoffs. Use how for runtime behavior.
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add cursor/plugins --skill whyOr 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/why ~/.claude/skills/whySafety box score
Grade B, low risk: 21 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 is a read-only investigation skill, but it reaches broadly across source control, issue trackers, docs, chat, observability, error tracking, and analytics via MCPs and `gh`. The main risk is breadth of remote data access and the fact that it explicitly tells the agent to use a broader agent mode instead of a constrained read-only mode.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 1 (low) | It directs the agent to run read-only history/PR commands like `git log --follow -p -- <file>`, `git blame -L`, `git show`, and `gh pr view`, but it ships no executable scripts or installers. |
| Network access | 2 (moderate) | It explicitly queries remote MCP-backed sources in parallel and uses `gh` against GitHub, so it will send search terms and code context to external services. |
| Filesystem reach | 0 (none) | The skill only reads repository history and metadata; there are no write/delete commands or paths outside the project. |
| Secrets & credentials | 0 (none) | No step asks for tokens, passwords, `.env` files, keychains, or browser sessions; it relies on whatever auth the environment already has. |
| Privilege & persistence | 0 (none) | There is no sudo, global install, service registration, shell startup editing, or OS/git configuration change. |
| Instruction hijack surface | 2 (moderate) | It explicitly says `Do not use readonly/Ask mode` and `Every spawn is generalPurpose with readonly: false`, which pressures the agent to leave a safer constrained mode. |
| Irreversible actions | 0 (none) | The skill is for investigation only; it does not push, deploy, publish, send messages, pay, book, or delete anything. |
| Transparency | 0 (none) | The shipped files are plain markdown prompts and playbooks; there are no binaries, blobs, or decoded-at-runtime code. |
Before installing
- Confirm which MCPs are actually enabled in your environment, since this skill will query any available source category it finds.
- Check whether your `gh` auth is read-only and scoped to the repos you expect, because the skill uses `gh pr view` and git history commands.
- Decide up front whether you want this skill to have access to chat, observability, Sentry, and warehouse data, since those are all in its normal search path.
Rated 2026-10-20 by an automated review of 13 files from the skill's repository.
Source
- Repository
- cursor/plugins (all skills from this repository)
- Path
- pstack/skills/why/SKILL.md
- Branch
- main
- Collection
- cursor
- Updated
- 2026-09-23
Related skills
- benchmark-due-diligence — Runs adversarial due-diligence on a benchmark the user envies — a founder, KOL, company.
- interrogate — Use for "interrogate", "adversarial review", "multi-model review", "challenge this", "stress test this code", "find blind spots", or "tear this apart".
- principle-attack-the-premise — Apply when two or more fixes that share one premise have failed the same gate.
- thermo-nuclear-code-quality-review — Run an extremely strict maintainability review for abstraction quality, giant files, and spaghetti-condition growth.
- differential-review — Performs security-focused differential review of code changes.
- comprehensive-research-agent — Ensure thorough validation, error recovery, and transparent reasoning in research tasks with multiple tool calls