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.
npx skills add kensaurus/cursor-kenji --skill audit-realworldgit clone --depth 1 https://github.com/kensaurus/cursor-kenjiWrote 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/skills/kensaurus/cursor-kenji/audit-realworld)<a href="https://agentmods.dev/skills/kensaurus/cursor-kenji/audit-realworld"><img src="https://agentmods.dev/badge/skills/kensaurus/cursor-kenji/audit-realworld/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/skills/kensaurus/cursor-kenji/audit-realworld"><img src="https://agentmods.dev/badge/skills/kensaurus/cursor-kenji/audit-realworld.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 3 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Tool Misuse · line 136 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
- high Tool Misuse · line 141 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
- high Tool Misuse · line 142 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
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 | $0.00058 | $0.03182 |
| Opus 5.5 | $0.00023 | $0.01273 |
| Sonnet 5.5 | $0.00012 | $0.00636 |
| Haiku 4.5 | $0.00006 | $0.00318 |
Grade A, and why
audit-realworld 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 — 288 lines — stays where its author put it; the contents beside it link to each section on GitHub.
audit-realworld — Full-Stack Gap Audit vs the RealWorld Reference
Degree of freedom: MIXED — Mode, parity, and routing [HIGH freedom];
Phase 4 contract-suite run [LOW freedom — run exactly] when a server or
suite exists.
Read-only. Findings and a prioritized gap report only — no code changes. Remediation goes to the skill named per finding.
RealWorld ("Conduit") is an objectively testable full-stack benchmark: formal
API spec, shared E2E suite, shared CSS theme, hosted API at
api.realworld.show.
The RealWorld API spec and its E2E suite are the source of truth — not any single community implementation. Reference repos vary; the spec and the passing suite do not.
RealWorld is a demo spec, not a production bar. It omits rate limiting, observability, caching, CI/CD, secrets, and deployment. Report feature/ pattern parity, then delegate hardening to
audit-security,plan-security-audit,plan-perf-audit,plan-rls-audit, andfull-stack-ship-discipline.
How to reason — Observe → Interpret → Classify → Severity
- Observe — quote the spec contract (or pattern) and the repo's
file:line(or suite result) - Interpret — does this layer implement the contract, or only the UI/route?
- Classify — Implemented / Partial / Missing / Diverges — or Bow-out if the repo is not a web CRUD+auth app
- Severity — core Conduit / authz / list-contract gaps = High; production-omitted concerns are routed, not scored as RealWorld misses
Worked example
Observe: Conformance mode.
GET /api/articlesreturns{ articles }with noarticlesCountand nolimit/offset(app/api/articles/route.ts). Bruno list request fails the count assertion. Auth header isToken <jwt>. Interpret: the list contract is partial; clients cannot paginate. Auth scheme matches the spec. Classify: Partial (pagination /articlesCount); auth Implemented. Severity: High — feed is a core Conduit surface. Finding: Article list | no pagination / articlesCount | High | complete-everything
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 Changed · +1 lines · -9 tokens per session 021aae6f3134
- 1mo ago First seen · 287 lines · 67 tokens per session scan A 859007c259f9
audit-realworld is a skill published in the GitHub repository kensaurus/cursor-kenji (9 stars, last pushed 26d ago), licensed MIT. It adds 58 tokens to every session and 3,182 once invoked, about $0.0002 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 2026-09-03.
Other skills, from other repositories
api-test-suite-generator
Generates comprehensive API test suites using Jest, Vitest, or Supertest from Express, Next.js, Fastify, or other API routes. Creates integration tests, contract tests, and edge case coverage. Use when users request "generate api tests", "create endpoint tests", "api test suite", or "integration tests for api".
api-mock-server
Creates mock API servers for testing and development with MSW, json-server, or custom handlers. Use when users request "API mocking", "mock server", "MSW setup", "test fixtures", or "mock API responses".
contract-testing-builder
Implements API contract testing to ensure provider-consumer compatibility using Pact or similar tools. Prevents breaking changes with contract specifications and bi-directional verification. Use for "contract testing", "API contracts", "Pact", or "consumer-driven contracts".
backend/testing-guide
A guide for writing backend tests: small unit tests, tests that check connected parts such as an API and database, and end-to-end tests that follow a complete user flow.
patrol-e2e-testing
Use when writing E2E/integration tests, testing native interactions like permissions or system dialogs, capturing UI regressions, or validating cross-platform behavior (Patrol 4.x).
testing
Use when writing or reviewing Flutter/Dart tests (unit, widget, golden), fixing flaky tests, adding coverage, or choosing between unit and widget tests.