Skill UI

Execute an implementation plan inline in the current session, task by task, using TDD and a durable ledger instead of per-task subagents. Choose it when your partner picks inline execution or no subagent tools exist. Covers briefs, rulings, completion contracts and final branch review.

This skill manages the end-to-end Navan expense reporting workflow. It allows users to retrieve detailed transaction data via API, automate multi-level approval workflows based on configured policies, and map expenses to specific General Ledger (GL) codes. It is designed to generate journal entry payloads for seamless synchronization with major accounting systems like NetSuite, Xero, and QuickBooks, streamlining financial reconciliation.

This workflow integrates with the Ramp API to manage corporate card transactions and expenses. It allows users to list, filter, and retrieve detailed transaction data, and crucially, sync these expenses directly to an accounting system (ERP). The process includes mapping transactions to general ledger accounts, making it ideal for finance teams handling expense reconciliation and accounting processes.

Atlas Ledger is a critical companion designed to provide lasting 'memory' to project contracts. It automatically processes detected deviations or 'drift' (e.g., from audits or reviews) and distills these instances into formal, actionable contract clauses (WHEN/DON'T/INSTEAD). These clauses are written to a persistent project memory file, Atlas.md. This mechanism ensures that lessons learned from failure are systematically captured, allowing future contract cycles to incorporate these rules, thereby strengthening overall project governance and reliability.

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.

A powerful tool for conversational time tracking. Users simply report time spent in plain language (e.g., "read papers 2h, gym 1h"), and the skill parses this into structured rows (Activity, Minutes, Date) within your Notion database. It ensures data integrity by flagging any uncertain details as 'To-confirm' rows and batch-asking questions, preventing data fabrication. Use it for logging daily activities or performing time reviews.

This advanced trading journal captures the full context of a trade, recording not only the fill price and P&L, but also the initial thesis, the defined plan (stop/target), and the emotional state at the time of entry. It integrates with Notion to log and review trades, grading the quality of decisions and execution against the user's own plan, ensuring users learn from process, not just profit.

This skill manages repeated rollouts and recursive prompting by maintaining an append-only decision ledger. It tracks trials, marks, and coherence checks to prevent auto-approval of live trading or destructive actions. Ideal for high-dimensional search, stochastic optimization, and ensemble comparison requiring visible evidence trails.

Runs a SHA-pinned integrity-forensics sweep on a paper (evidence ledger, auditor proposals, rules-only report), converting verdicts into typed policy gates (BLOCK/WARN/NO_NEW_BLOCKER) and an obligations ledger. Used for pre-submission integrity checks, peer-review forensics and resubmission pipelines.

This guide instructs authors on how to package theoretical proofs for submission to COLT (Conference on Learning Theory). Since COLT does not have a formal artifact track, it emphasizes treating the proof appendix as the primary 'artifact.' It details best practices for creating a verifiable proof ledger, rigorously tracking constants and external results, and managing supplementary code to ensure maximum academic credibility.

Provides a comprehensive guide for authors responding to highly technical conference reviews, particularly focusing on the critical 'one-shot major revision response letter.' It details strategies for maintaining strict double-anonymity, structuring the rebuttal around decision-critical points (soundness, reproducibility), and creating a detailed change ledger to systematically address every minimum necessary change requested by reviewers.

A comprehensive guide for hardening deep learning papers, particularly in Computer Vision (CVPR). It details best practices for achieving rigorous reproducibility by utilizing the Compute Reporting Form (CRF), maintaining a detailed recipe ledger, ensuring benchmark split hygiene, and meticulously documenting all training hyperparameters, seeds, and variance analyses for academic publication.