Skill UI

This specialized tool generates a pure LaTeX abstract for academic papers. It takes the fully revised body, core topic, and detailed preferences (such as audience, style, and emphasis) as inputs. It ensures the abstract is a concise, faithful summary, covering the problem, approach, evidence, and main results while adhering to strict structural and stylistic guidelines.

This stub experiment generator is designed for the meta-paper-write demo. It takes a topic phrase and deterministically produces a 20-row CSV file containing baseline and generated results. This ensures stable and reproducible results for demonstration purposes, but should not be used for real scientific data.

This tool extracts and standardizes comprehensive paper-writing preferences from a user's initial request. It determines whether the generation should proceed directly using safe academic defaults or if it needs to initiate a preference interview, ensuring all structured data required for robust academic drafting is captured.

This specialized tool drafts a single, academically rigorous section of a research paper. It accepts detailed structured inputs, including a complete writing plan, overall paper outline, citation budget, and optional figure/table context. The output is a pure LaTeX fragment that adheres to strict structural conventions for various sections (e.g., Introduction, Method, Results). It ensures the section advances the paper's single narrative thread while maintaining professional academic quality.

Use PaperJSX to automate the creation of professional documents, including PPTX presentations, DOCX reports, XLSX spreadsheets, and PDF files. By providing a structured JSON specification, this tool generates ready-to-use files, eliminating manual formatting and ensuring consistency across large-scale reporting and content generation pipelines.

A robust toolkit for deterministic, structural PDF manipulation. This skill allows users to perform programmatic tasks such as extracting text and tables from reports, merging multiple PDF files, splitting documents by page ranges, filling out form fields, or generating entirely new PDFs from structured data (e.g., JSON). It is designed for explicit, byte-level operations, distinguishing it from LLM-based natural language rewriting tools.

Use the Composio CLI to automate the entire Pull Request review and Continuous Integration (CI) triage process for GitHub and GitLab. This tool fetches code diffs, analyzes failure logs, posts detailed inline review comments, and iteratively patches and re-runs CI checks until the code passes, significantly enhancing code quality and developer efficiency.

This tool automates the diagnosis of production errors tracked by Sentry. By leveraging the Composio CLI, it retrieves comprehensive issue details, full stack traces, breadcrumbs, and associated commits. It maps stack frames to local source code, allowing agents to propose fixes and triage issues directly, eliminating manual copy-pasting.

An internal tool designed to validate candidate meta-skills (SKILL.md) before they are registered in the system. It performs comprehensive linting checks, including structural validation (G1: parsing, reference checks, XML escaping) and a scheduler dry-run (G2) using stub executors, ensuring the skill is deterministic and functional. Output diagnostics in JSON format.

An internal utility designed to manage the entire lifecycle of meta-skill proposals. It provides atomic functions to write new skill proposals, list candidates for review, and formally accept validated proposals into the active skill set. This tool is crucial for automated skill pipeline development and version control.

An internal, non-user-invocable tool designed to validate candidate meta-skills. It runs critical G3 (positive) and G4 (negative) smoke tests by simulating meta-resolution using LLM-generated fixtures. This ensures the structural integrity and robustness of the skill in a cross-vendor environment.

A utility for managing and installing specialized AI skills into the Codex environment. It supports fetching curated skills from predefined sources (like openai/skills) or installing custom capabilities directly from any GitHub repository, including private ones. It handles listing available skills, installing specific modules, and provides robust fallback methods (download/git sparse checkout).