Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/emaraschio/cursor-commandsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/emaraschio/cursor-commands/scoped-audit)<a href="https://agentmods.dev/commands/emaraschio/cursor-commands/scoped-audit"><img src="https://agentmods.dev/badge/commands/emaraschio/cursor-commands/scoped-audit/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/emaraschio/cursor-commands/scoped-audit"><img src="https://agentmods.dev/badge/commands/emaraschio/cursor-commands/scoped-audit.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00031 | $0.00638 |
| Opus 5.5 | $0.00012 | $0.00255 |
| Sonnet 5.5 | $0.00006 | $0.00128 |
| Haiku 4.5 | $0.00003 | $0.00064 |
Grade A, and why
scoped-audit scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 44 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Overview
Run a large "check-N-things" task (audit, inventory, or migration) as a scoped, plan-first workflow: decompose the surface, fan out capped parallel subagents, verify findings before reporting, and make changes only after explicit approval. Full workflow: .cursor/skill-contracts/scoped-audit/SKILL.md (user install: ~/.cursor/skill-contracts/scoped-audit/SKILL.md).
Defaults
| Setting | Default |
|---|---|
| First pass | Scoped sample only (never the full surface) |
| Concurrent subagents | Capped (≤ 6) |
| Changes | None until the plan is approved |
Steps
- Read
.cursor/skill-contracts/scoped-audit/SKILL.mdfor the full agent contract; if that file is missing, read~/.cursor/skill-contracts/scoped-audit/SKILL.md. - Execute phases in order (Scope → Plan → Approval gate → Scoped sample → Parallel fan-out → Verify → Report → optional gated changes); do not skip the approval gate or the verification step.
- Report the catalog of verified findings; propose changes separately and wait for approval before editing.
Anti-patterns
- Sample before you fan out. Trigger: starting a large check-N-things pass. Wrong: running the full surface before the approach is validated. Correct: run a scoped sample first to confirm the finding shape and verification method. Reason: an unvalidated full run across a large surface burns tokens and compounds mistakes.
- Verify findings before reporting. Trigger: assembling the catalog from subagent results. Wrong: reporting a subagent summary as fact without rechecking. Correct: re-check each finding against the source and flag what cannot be confirmed. Reason: large parallel passes are where agents hallucinate, so an unverified catalog is worse than none.
- Make no changes until the plan is approved. Trigger: the user asks to audit and fix in one pass. Wrong: editing or migrating across the surface before approval. Correct: present proposed changes separately and wait for an explicit approval gate. Reason: bulk edits before review are hard to unwind across a large surface.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 44 lines · 31 tokens per session scan A b79bae614be7
scoped-audit is a command published in the GitHub repository emaraschio/cursor-commands (9 stars, last pushed 5d ago), licensed MIT. It adds 31 tokens to every session and 638 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
ship-and-observe
Take merged, green code to a verified, monitored production release — deploy, confirm the revision is live, smoke-test, observe, and roll back if needed.
deadcode
Find dead code with a configured Knip baseline, then delete it by category behind a ratchet.
refactor
Systematic refactoring: analyze → split → extract → verify behavior, with no truncation or omitted lines.
arch-boundaries
Install mechanically-enforced architecture boundaries that fail CI on spaghetti imports.
auth-flows
Read-only audit of app-layer auth — route×gate matrix, getSession vs getUser, middleware-as-only-gate.
housekeep-gates
Consolidate accreted CI gates into one aggregator required check — apply phase by phase.