Results for “premium-model-leakage”

Test LLM applications for system prompt leakage using manual payloads, garak, and Promptfoo to extract embedded secrets and routing logic.

Systematically extract hidden system prompts, core directives, and invisible context intentionally concealed within Large Language Model (LLM) applications. This skill utilizes targeted linguistic engineering and boundary manipulation to bypass prompt opacity.

Detect model stealing, model inversion, and membership inference performed through inference-API abuse by monitoring query patterns, applying output perturbation, and red-teaming your own model's extractability.

AI/ML security playbook. Use when assessing model supply chain attacks (pickle RCE, poisoned weights), adversarial examples, model poisoning, model stealing, data privacy attacks (membership inference, model inversion), and autonomous agent security risks.

Read-only OWASP LLM Top 10 audit of app-facing AI features: prompt injection, data leak, supply chain, poisoning, unsafe output, excessive agency, system-prompt leak, RAG/embedding risks, misinformation, unbounded consumption. Use when "audit LLM security", "prompt injection", "jailbreak my chatbot", "is my AI safe".

Generates a single-frame HTML template with a cinematic film-leak aesthetic: warm light leaks, 35mm grain, 2.39:1 letterbox, and serif typography for opening titles or chapter cards.

Detect poisoned training data and backdoored models across the ML pipeline using statistical analysis, activation clustering, and spectral signatures.

Search and browse Civitai models, images, and videos; extract generation prompts; and get download commands for LoRAs and checkpoints.

Detect and prevent public-over-private name resolution in npm, PyPI, and Maven dependency manifests.

Tests APIs for excessive data exposure where endpoints return more data than the client application needs, relying on the frontend to filter sensitive fields. Maps to OWASP API3:2023 Broken Object Property Level Authorization.

Least privilege is easy to state and hard to hold: grant each identity only the

Documents MariaDB-specific GRANT behavior, including privilege scope levels, implicit account creation, MariaDB-only privileges, roles, proxy grants, resource limits, and TLS options, to help write accurate GRANT statements.

Elevate from a low-privilege user account to root access on a compromised Linux system by exploiting misconfigurations, vulnerable services, kernel exploits, and weak permissions.

Execute and analyze AI Data Poisoning attacks. By subtly injecting malicious or targeted misinformation into an LLM's training or fine-tuning dataset, an attacker can covertly manipulate the model's future outputs, implant backdoors, or enforce biases without altering the model architecture.

Performs STRIDE-A threat model analysis of repositories and systems, producing architecture overviews, DFD diagrams, prioritized findings, and executive assessments. Supports both single analysis and incremental updates with change tracking.

Deploy open models or custom weights from Model Garden to Agent Platform endpoints, check deployment status, verify serving endpoints, or clean up resources by undeploying models and deleting endpoints.

Use to design VIP or tiered support programmes and to test whether higher tiers actually deliver better outcomes. Trigger for "VIP support programme", "service tiering", queue priority for premium customers, entitlement vs priority, tier inflation, or "does our enterprise tier get faster support".

VIXEN v1.0 — Dual-mode emerging movers scanner. Built from FOX v1.6 live trading data (+34.5% ROI). Two entry modes: STALKER (steady accumulation, score 6+, catches the SM buildup BEFORE the explosion) and STRIKER (violent FIRST_JUMP, score 9+, with raw volume confirmation to filter blow-off tops). 2-hour per-asset cooldown after Phase 1 exits. DSL High Water Mode mandatory.

Detect and defend against prompt injection hidden in documents, web pages, and images consumed by an agent.

Scan Model Context Protocol servers and tool metadata for poisoning, SSRF, and unauthenticated exposure.

Documents the Bugatti brand design system: a black-and-white interface with custom typefaces, no accent color, and full-bleed photography as the only visual element.

Design leakage-safe feature engineering strategies for tabular/time-series datasets. Use when: (1) preparing model-ready features, (2) selecting transformations and encodings, (3) documenting feature lineage. NOT for: model serving or infra provisioning.

Exploit privileged pods, host mounts, runC CVEs, and exposed Docker sockets to break out of a container and reach the underlying host during authorized container-security assessments.

Chief Security Officer mode. (gstack)

Identify sensitive data exposure vulnerabilities including API key leakage, PII in responses, insecure storage, and unprotected data transmission during security assessments.

Audit agent skills, plugins, prompts, manifests, scripts, dependencies, and bundled assets for provenance, prompt-injection, permission, execution, exfiltration, persistence, and update risk. Use when evaluating a third-party skill before installing, enabling, updating, publishing, or distributing it; reviewing an untrusted SKILL.md, agent configuration, MCP integration, archive, or repository; comparing a package with a known-good version; or investigating unexpected tool, network, credential, or filesystem behavior.

Retrieves upcoming earnings announcements for US stocks using the Financial Modeling Prep API, filtered by market cap over $2B, and organizes them into a markdown table grouped by date and timing.

Use Pacu modules for AWS privilege escalation, persistence, and backdooring during authorized penetration tests.

Identify AWS IAM privilege escalation paths by analyzing policies for dangerous permission combinations and least-privilege violations using boto3 and Cloudsplaining-style analysis.

Detect and respond to Adversary-in-the-Middle (AiTM) phishing attacks that use reverse proxy kits like EvilProxy, Evilginx, and Tycoon 2FA to bypass MFA and steal session tokens.

Platform APIs for model management, pricing, and usage tracking

Crea un sistema de diseño consistente con múltiples temas (light/dark/custom), Material 3, tipografía personalizada y cambio dinámico de temas en Flutter.

设计原型专家团入口。用于 Codex CLI 的 $expert-design 调用。 覆盖需求发现、设计系统选择、原型生成、质量审查和导出交付。 触发词：设计原型、UI、UX、设计系统、高保真、HTML 原型、品牌设计、页面设计

PR Review Expert

Use to check whether a conversation export, dataset or AI pipeline is leaking personal data that should have been redacted, and to measure how well the redaction actually works. Trigger for "is this export safe to share", "check our redaction", "can we use support transcripts for training", sending transcripts to a vendor or model, "is there PII in this dataset", or before opening support data to a wider audience.