Browse Skills

Jailbreak Override is a skill that attempts to remove safety restrictions from an AI agent by instructing it to ignore content policies and hide that behavior.

MCP Atlassian Hint Injection demonstrates how a crafted hint parameter causes an MCP Atlassian server to leak its Authorization header. Use it to test for CVE-2026-27825/27826 prompt injection vulnerabilities.

MCPwn CVE-2026-33032 demonstrates the MCPwn runaway invocation pattern that can trigger CVE-2026-33032. Use it to test whether an MCP client can be coerced into retrying a tool call indefinitely until a target API key is returned.

Prompt Injection Test provides a test skill containing prompt injection patterns to evaluate LLM agent defenses against malicious instructions.

Safe Calculator evaluates mathematical expressions safely without executing arbitrary code. Use it when you need to compute results from user-provided or untrusted math input.

Safe File Reader reads files from a documents directory with validation so an agent can access only allowed files. Use it when an agent needs to read document contents without exposing arbitrary files.

Simple Formatter formats text according to specified style guidelines, including sentence capitalization, spacing, and punctuation. Use it when raw text needs cleanup before publishing or further processing.

Simple Math performs basic arithmetic operations safely using standard operators without eval or exec. Use it when an agent needs to calculate addition, subtraction, multiplication, or division with input validation.

Weather Assistant retrieves current weather conditions and forecasts for any location worldwide. Use it when you need real-time weather, rain predictions, or multi-day outlooks for a city or region.

Patch Creation documents how to create and register new patches for tweakcc, covering pattern matching, config wiring, registry updates, and UI integration.

Install Anti-Slop adds and configures the anti-slop Oxlint plugin in a local TypeScript or JavaScript repository. Use it when adding anti-slop lint rules, copying the plugin, configuring opinionated Oxlint rules, or migrating an existing local anti-slop setup.

Ni Haixia equips an AI agent with the complete traditional Chinese medicine knowledge and clinical reasoning of Ni Haixia, covering six-meridian syndrome differentiation, classical formulas, acupuncture, and herbal knowledge. Use it when you need TCM insights or to emulate Ni's teaching style.