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 jaktestowac/awesome-copilot-for-testers --skill testing-llm-guardrailsgit clone --depth 1 https://github.com/jaktestowac/awesome-copilot-for-testersWrote 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/jaktestowac/awesome-copilot-for-testers/testing-llm-guardrails)<a href="https://agentmods.dev/skills/jaktestowac/awesome-copilot-for-testers/testing-llm-guardrails"><img src="https://agentmods.dev/badge/skills/jaktestowac/awesome-copilot-for-testers/testing-llm-guardrails/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/jaktestowac/awesome-copilot-for-testers/testing-llm-guardrails"><img src="https://agentmods.dev/badge/skills/jaktestowac/awesome-copilot-for-testers/testing-llm-guardrails.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 YARA Match · line 2 YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).Fix: Remove offensive tool references and exploit code. Legitimate agent skills should not contain penetration testing tools, exploit frameworks, or reconnaissance utilities.
- high Prompt Injection · line 49 This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.Fix: Remove or rewrite any text that instructs the agent to ignore prompts, override safety rules, or trust unverified content. Ensure skill content cannot be injected to alter agent behavior.
- high Prompt Injection · line 134 This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.Fix: Remove or rewrite any text that instructs the agent to ignore prompts, override safety rules, or trust unverified content. Ensure skill content cannot be injected to alter agent behavior.
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 | $0.00109 | $0.02653 |
| Opus 5 | $0.00055 | $0.01326 |
| Sonnet 5 | $0.00022 | $0.00531 |
| Haiku 4.5 | $0.00011 | $0.00265 |
Grade B, and why
testing-llm-guardrails scanned grade B with 1 finding 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 12d 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.
Instruction-override phrasingmediumPrompt injection
Text telling the model to disregard its earlier instructions or safety rules is the shape of a prompt injection, whoever wrote it.
- **A corpus, not a handful of tricks.** "Ignore previous instructions" is one case. Categories, versioned and grown from real attempts, is a suite. Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.
How it starts
The opening of the file, as written. The whole thing — 165 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Testing LLM Guardrails
Use this skill when an LLM feature can affect something - render output to a user, call a tool, write to a database, send a message - and you need to know what happens when the model is manipulated or simply wrong.
Two layers, tested together because they fail together:
- Guardrails - the runtime validation between the model and the world: schema, moderation, PII filtering, refusal handling, tool authorisation. Deterministic code you can unit-test.
- Adversarial resistance - whether those guardrails hold when someone tries to defeat them, including through content the model reads rather than what the user types.
The principle that drives all of it: the model is untrusted input, and so is everything it reads. A prompt is not a security boundary. Code around the model is.
Authorization Gate - read before anything else
Adversarial testing is security testing. Before running a single case, confirm in writing:
- the system is yours, or you have explicit written permission from its owner
- the target environment is named and non-production, or production testing is explicitly authorised
- the time window and rate limits are agreed
- who to notify on a real finding, and how findings are handled
If any of that is missing, stop and say what is needed. Do not "just try one". The same rule as testing-application-security, for the same reason.
Scope boundaries that always apply: test only the surfaces you were authorised for, do not exfiltrate real user data as proof, never use another tenant's live data as a test payload, and report findings privately before they go anywhere public.
When to Use
- an LLM feature is about to reach real users
- the model can call tools, especially any that mutate state or spend money
- the feature reads content it did not author: retrieved documents, user uploads, web pages, webhook payloads, tool results
- model output is rendered as HTML, markdown, or executed as code or SQL
- a security review asks whether the AI feature can be manipulated
- an incident involved the model doing something it should not have
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 165 lines · 109 tokens per session scan B 5f65d6fee165
testing-llm-guardrails is a skill published in the GitHub repository jaktestowac/awesome-copilot-for-testers (113 stars, last pushed 16d ago), licensed MIT. It adds 109 tokens to every session and 2,653 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it B with 1 finding (instruction-override phrasing). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
fix-tests
Focus on all unit + command tests (pytest --exclude tests/integration). Make sure they pass and fix errors. If you run into anything very odd: stop, and let me know. Mutate test code first and let me know if you think you should update application code.
character-animation-qa
Review local character animation with schema checks, Playwright browser previews, frame sampling, and FFmpeg/ffprobe final output checks.
migrate-xunit-to-xunit-v3
Migrate .NET test projects from xUnit.net v2 to xunit.v3 and fix v3 breaks. Use for package/CPM conversion, OutputType=Exe, preserving the VSTest or MTP runner (including projects currently using YTest.MTP.XUnit2), incompatible TFMs, async void tests, string-to-Type attributes, custom Fact/Theory/BeforeAfterTest…
nunit
Write, run, or repair .NET tests that use NUnit. Use when a repo uses NUnit, [Test], [TestCase], [TestFixture], or NUnit3TestAdapter for VSTest or Microsoft.Testing.Platform execution. USE FOR: writing or reviewing NUnit tests; using [Test], [TestCase], [TestFixture], [SetUp], [TearDown] attributes; configuring…
crap-score
Calculates CRAP (Change Risk Anti-Patterns) for a named .NET method, class, or file. USE FOR: explicit CRAP calculation or coverage-and-complexity risk within that named target, including which tests to prioritize. DO NOT USE FOR: project-wide coverage/CRAP, plateaus, or project-wide blockers/priorities…
michel-create-packmind-dataset
Seed a local Packmind instance with a realistic dataset — one organization populated with standards, commands, and skills — so an autonomous agent can exercise its own changes against lifelike data instead of an empty app. Use this whenever you need populated Packmind data to verify a change end-to-end: reproducing a…