Browse Skills

Use when auditing an AAAI main technical track submission for OpenReview readiness, double-blind anonymity, page limits, reproducibility checklist, supplementary material, author limits, multiple-submission policy, and AAAI AI-use policy compliance.

Audit and design experiments for AAMAS papers, ensuring interaction claims are tested via self-play, population sweeps, and deviation tests, with proper metrics, replicability, and ablation structures.

Provides an evidence map and checklist to ensure AAMAS papers have reproducible interaction claims, covering game definitions, opponent sets, seeds, compute, and uncertainty. Use before submission and camera-ready.

Strategic guidance for the AAMAS peer review process, covering rebuttal tactics, scoring dimensions, and reviewer expectations for multi-agent and game theory papers.

Audits AAMAS conference submissions for OpenReview completeness, page limits, double-blind anonymity, track selection, supplementary material rules, and desk-reject triggers before the deadline.

Reviews animation and motion code to ensure high craft quality, checking easing, performance, accessibility, and physical correctness. Use it when you need a strict motion design reviewer.

Reviews animation and motion code against a high craft bar, flagging issues with easing, timing, performance, and accessibility. Use when you need rigorous motion design review.

Detects prompt injection attacks targeting LLM-based applications using a multi-layered defense combining regex, heuristics, and a DeBERTa classifier. Use it to scan user inputs before forwarding to an LLM, audit logs, or during security assessments.

Tests machine-learning model files and model-loading services for remote code execution vulnerabilities caused by insecure deserialization, archive traversal, Hydra instantiation, and memory corruption. Use during authorized penetration testing of ML pipelines.

Tests LLM-backed apps, chatbots, and agents for prompt injection, jailbreaks, and system prompt leaks using direct, indirect, and encoding techniques.

Detect and test for OWASP API3:2023 Broken Object Property Level Authorization (BOPLA) vulnerabilities, including excessive data exposure and mass assignment, with automated scanning.

Tests APIs for SQL, NoSQL, command, LDAP, and SSRF injection vulnerabilities. Use when assessing API endpoints that accept user input for database queries, system commands, or external requests.