Skill UI

Design and revise company commercial policy: data-backed discount matrices, approver thresholds, exception flows, and deal framework, plus governance linting for contradictions, gaps, and cliff edges.

Routes commercial inquiries to one of seven sub-skills covering pricing strategy, deal desk approvals, partnership economics, channel mix, commercial policy, RFP responses, and bookings forecasts. Returns concise digests with cited best-practice challenges.

A powerful tool for deterministic and reproducible retrieval of evidence from specialized academic and policy corpora. It supports searching arXiv (preprints), PubMed/PMC (biomedical literature), and US policy documents. Users can check corpus freshness, retrieve stable metadata, and fetch full-document Markdown for high-reliability research tasks.

A powerful meta-skill that enforces rigorous security compliance by running three independent gates on a proposed operation: policy/governance review, secret/credential scanning, and audit log emission. It uses a strict arbitration rule (DENY > WARN > ALLOW) to provide a comprehensive safety verdict for code changes or environment manipulations. Ideal for enforcing high-security coding standards.

This skill provides a highly rigorous framework for analyzing complex public issues, platform governance, and institutional compliance. It is designed to structure evidence boundaries, assess procedural justice, and track public commitments by synthesizing source ledgers and multiple data points. It is intended as a specialized sub-routine triggered by the main CrossFrame suite, not for standalone use. Key focus areas include source verification, identifying weak signals, and mitigating AI compliance performance risks.

This skill guides the creation of a comprehensive Service Level Objective (SLO) document. It helps define Service Level Indicators (SLIs) by identifying critical user journeys and measuring key metrics like success rate and latency. It calculates the allowed "error budget" and establishes a burn rate policy to maintain service reliability while enabling feature velocity and principled trade-offs.

Implements a comprehensive defense-in-depth framework for autonomous AI agents that interact with external tools. It mitigates risks like prompt injection or tool misuse by enforcing strict tool allowlisting, binding tools to least-privilege identities, requiring human-in-the-loop approval for high-impact actions, and utilizing policy enforcement (like NeMo Guardrails) to bound the blast radius and ensure compliance.

This skill mints tamper-evident, post-quantum-signed receipts for consequential agent actions (e.g., deployment, payment, access grant). It records who, what, and under which policy the action occurred, ensuring the evidence cannot be altered later. By providing verifiable proof from the certificate alone, it is crucial for meeting regulatory compliance requirements, such as the EU AI Act, and building auditable, trustworthy systems.

This skill provides expert capabilities for training and fine-tuning transformer language models using the TRL (Transformers Reinforcement Learning) library. It supports state-of-the-art post-training methods, including Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and Reward Model training. Use this skill to align and customize foundation models for advanced tasks via CLI commands.

KubeEye is a comprehensive cluster inspection tool designed for Kubernetes environments, particularly within KubeSphere. It systematically detects issues across workloads, nodes, and configurations by integrating multiple policy engines, including OPA/Rego, PromQL, file integrity checks, kernel parameter validation, and systemd health checks. Users can define custom inspection rules (InspectRule), schedule checks (InspectPlan), and retrieve detailed reports (InspectResult) to ensure cluster compliance and stability.

A structured presentation deck designed for rigorous review of safety incidents, security risks, and policy compliance. It utilizes distinct visual cues, including hazard stripes, L1/L2/L3 severity tiers, and dedicated code blocks for policy review (e.g., YAML). Ideal for red team exercises, post-incident reviews, and high-stakes technical briefings.

This skill is designed to mitigate false-positive safety triggers from strict AI classifiers (e.g., cyber, bio/chem, reasoning). It operates by making minimal, surgical edits to sensitive phrases, abstracting away specific domains or actions while preserving the original benign intent and structure. It ensures the prompt remains policy-compliant without losing its core purpose, making it ideal for academic or defensive inquiries.