dingxingdi
- 389 skills
- 0 followers
- 20 hours ago last updated
- ▌ Autonomous Web Data Collection Engineering · dingxingdi bundleUse this skill when the user wants to autonomously gather datasets from open web sources by writing and executing scrapers or API harvesters. It is triggered by natural language requests such as 'collect all papers from this year's top AI conference', 'give me a CSV of NBA player stats for 2023', 'download daily stock prices for S&P 500 companies', or 'create a dataset of children's book abstracts from the web'. Use this when the agent needs to plan a data acquisition strategy, identify stable web anchors or API endpoints, and generate executable Python scripts to build a curated dataset from scratch.
- ▌
- ▌ Landmark Grounded Semantic Exploration · dingxingdi bundleSkill: landmark-grounded-semantic-exploration
- ▌ Hierarchical Rubric Based Artifact Grading · dingxingdi bundleUse this skill when you need to evaluate complex professional artifacts—such as physics proofs, financial investment memos, chemistry reports, or consulting strategies—against expert-level, multi-dimensional rubrics. It is triggered by requests like 'grade this professional report using expert criteria', 'apply weighted rubrics to this PhD-level paper', 'evaluate this MBA memo for reasoning and extraction quality', or 'use a locked checklist to score professional knowledge'. Ordinary users might say 'check if the physics math is right', 'see if the business plan covers all the risks', or 'score this report based on the official expert rules'.
- ▌ Text Data Privacy Sensitivity Adjudication · dingxingdi bundleUse this skill when you need to evaluate how much personal or private information is contained within a specific text. It is triggered by requests like 'Is this email too personal to share?', 'Check if this blog post has private stuff in it', 'How sensitive is this medical question?', 'Does this text reveal who wrote it?', or 'Rate this post from harmless to extremely private'. It helps ordinary users understand if their writing contains 'named entities' (like names/phones) or 'sensitive topics' (like health/legal issues) that could lead back to them.
- ▌ Multi Agent Belief Driven Coordination · dingxingdi bundleUse this when the user wants planning data for scenarios where an agent must track what other actors know, believe, or intend (Theory of Mind). Trigger it for requests like 'make tasks where the agent has to guess what its partner knows', 'solve a coordination problem with limited communication', 'plan a strategy that uncovers a deception or forms an alliance', or 'generate unexploitable strategies for competitive games with hidden info where you must exploit an opponent's blind spots.'
- ▌ Causal And Counterfactual Video Reasoning · dingxingdi bundleUse this skill when the user asks questions like 'is this a crime?', 'is there an anomaly here?', 'is this person acting suspicious?', 'why did the fight break out?', 'is this area safe?', 'classify this dangerous event', or 'pinpoint the exact moment the robbery started'. Trigger it when the agent must distinguish between normal behavior and anomalous/illegal activities, deduce the hidden motives of actors in a conflict, or evaluate the causal progression of crimes and accidents in surveillance or open-world videos.
- ▌ Autonomous Repository Exploration And Reuse · dingxingdi bundleUse this skill when you need to understand, deploy, adapt, or wrap an unfamiliar code project or third-party GitHub repository. It is perfect for requests like 'I'm totally lost in this GitHub code, help me use it', 'deploy this AI research project', 'run the training and inference for this paper repo', or 'wrap this research code into a simple reusable Python tool'. Trigger it when an agent must explore deep folder structures, resolve missing weights/dependencies, follow tangled code connections, and execute end-to-end research pipelines (Setup, Download, Training, Inference, Evaluation) or build functional wrappers based on execution feedback.
- ▌ Interactive Underspecified Issue Resolution · dingxingdi bundleUse this skill when the user provides a software bug report or feature request that is vague, missing error logs, or lacks specific file locations. Trigger this for requests like 'fix the crash in the login module' (without giving the log), 'the update button doesn't work', 'make the questions a bit clearer', or 'there is an error in the data processing part'. This skill is essential when you feel you are guessing about where the bug is or what the expected behavior should be; it guides you to detect the ambiguity and ask the user for the missing details before you start writing code.
- ▌
- ▌
- ▌
- ▌ Focus Controlled Element Aware Summarization · dingxingdi bundleUse this skill when the user wants a summary controlled by a strict lens, specialized constraint, or a specific audience register (e.g., layman translation). Trigger it for requests like 'cover only the who/when/what,' 'focus strictly on high-level goals,' 'explain this complex report in simple terms,' 'write a news summary covering the 5W elements,' or 'keep the summary under 700 characters in a bulleted list.' This is especially useful for maintaining coverage, ensuring safety omissions are respected, and shifting technical registers for non-expert readers.
- ▌ Exploration Guided Unfamiliar App Adaptation · dingxingdi bundleUse this skill when the user wants data where the agent must first learn an unfamiliar app's layout, handle GUI changes from version updates, or move tasks between mobile apps and websites using exploration. Trigger it for requests like 'the app updated and the button moved, can you find it?', 'make it figure out a new app by looking around first,' 'use this new store like you used the old one,' or 'go to the website and do what you did on the app.' This skill evaluates whether the agent can convert exploration evidence into app-specific navigation priors to overcome cross-version, cross-platform, or entirely new UI environments without relying on stale, memorized habits.
- ▌
- ▌
- ▌
- ▌
- ▌
- ▌ Recursive Hierarchical Aggregation · dingxingdi bundleUse this skill when the user provides a request involving a massive amount of data, logs, or dense, high-resolution artifacts (like a giant technical diagram) that are far too large or complex for a single context window. It is triggered by layman requests like "summarize all these hundreds of logs," "break this complex map down into bits," "make a summary of summaries," or "digitize this entire layout piece by piece." Trigger it for phrases such as "it's too much data to read," "get a bird's-eye view," "scan all the elements across the whole project," or "chunk everything up and tell me the main points."
- ▌ State Centric Action Orchestration · dingxingdi bundleUse this skill when the user wants an agent to handle complex projects by bridging different 'expert sub-tasks', tracking deeply nested historical logs to see what previously changed, or utilizing a shared memory blueprint. It is highly applicable when addressing a massive, unorganized environment where the agent posts a broadcast request (blackboard pattern) to asynchronously share discoveries, or needs to traverse a chronological history of prior actions. Trigger phrases include “keep track of what we've found so far,” “let the different tools see each other's work,” “tell me where you moved the items earlier,” “manage a project board for all the workers,” or “look at your action history to figure out the current status.”
- ▌ Symbolic Reactive Replanning And Repair · dingxingdi bundleUse this when the user wants planning data testing an agent's ability to update its course of action when the underlying structure or state of the world changes unexpectedly or when current actions become unfeasible. Trigger it for requests like 'make a task where the plan breaks halfway through', 'detect why I can't do the next step', 'verify if a subtask worked', 'perform a pre-flight check on the next action', 'fix a plan that got stuck', or 'generate data for reacting to changing maps or missing items during execution'.
- ▌ Contextual Privacy Preserving Disclosure · dingxingdi bundleUse this skill when the user wants the agent to write a message on their behalf and the environment contains private details that may or may not be appropriate to share. Trigger it for requests like “draft the weekly update,” “reply to my manager,” or “send this recap,” especially when the agent has access to calendars, notes, chats, or other personal records. It is the right skill when the challenge is to stay helpful without leaking something that technically appears in context but should not be forwarded.
- ▌ Epistemic Vigilance And Source Integrity · dingxingdi bundleUse this skill when you need to write or verify a document where every claim MUST be anchored to a specific, ID-verified source and checked for boundary-condition accuracy. Trigger it for requests like “verify these technical findings,” “ensure no fake citations are included,” “check the boundary conditions of this device design,” or “map every claim to a retrieved span.” It is critical for enforcing ‘closed-world citation’ (abstaining rather than guessing) and auditing the formal alignment between retrieved evidence and final claims.
- ▌ Autonomous Data Lake Discovery And Retrieval · dingxingdi bundleUse this skill when the user provides a data science question that requires finding the right files within a large and messy 'data lake' rather than having the data pre-specified. It is triggered by layman requests such as 'find the right data to answer this in our lake', 'search our huge project for information about X and analyze it', 'there are thousands of files, pick the relevant ones to solve Y', or 'I don't know which CSV has this info, please find it and calculate Z'. It is specifically for handling scale and heterogeneity in data environments where discovery is the first crucial bottleneck.
- ▌ Autonomous Software Performance Optimization · dingxingdi bundleUse this skill when you need to improve the runtime efficiency (latency) or resource footprint (memory) of an existing script or codebase through iterative self-improvement. It is triggered by natural language requests like 'make this run faster', 'optimize the memory usage of this function', 'improve the efficiency score relative to human solutions', or 'generate code that outperforms the standard compiler output'. It guides the agent in using empirical feedback from a sandbox—ranging from high-level algorithmic refinement to low-level assembly superoptimization—to progressively reach peak computational performance.
- ▌
- ▌
- ▌ Code Vulnerability And Security Adjudication · dingxingdi bundleSkill: code-vulnerability-and-security-adjudication
- ▌ Evidence Synthesis And Insight Derivation · dingxingdi bundleUse this skill when the writing task requires moving beyond a simple list of facts to produce reasoned insights, justify a verdict, or synthesize massive amounts of fragmented evidence. Trigger it for requests like 'write an investigative report,' 'analyze the trends,' 'give me a 360-degree view of these 100 documents,' 'explain why this claim is false based on the evidence,' or 'draft a legal reasoning section.' It perfectly resolves tasks where raw data must be explicitly mapped to overarching conclusions, truths, or rules.
- ▌ Idiosyncrasy Removal And Stylistic Polish · dingxingdi bundleUse this skill when the writing needs to move beyond simple 'AI fluency' to achieve expert-level literary quality, human personality, and varied vocabulary. Trigger it for requests like 'make this sound like a real person,' 'add more vivid sensory details,' 'remove the robot-like cliches,' 'make the ending echo the beginning,' or 'don't use the same words over and over.' It is specifically designed to purge 'AI slop' (formulaic metaphors, low-density adjectives, repetitive sentence loops) while injecting high-resolution imagery, diverse vocabulary, and professional structural closure.
- ▌ Vulnerability Reproduction And Poc Generation · dingxingdi bundleUse this skill when the user wants to generate a Proof-of-Concept (PoC) or exploit script to trigger a specific memory safety or logic vulnerability in a codebase. Trigger this for requests like 'create a crashing input for this buffer overflow', 'generate an exploit that triggers this ASan report', 'find a PoC for the use-after-free bug', or 'reproduce this memory leak in the C++ repo'. It is explicitly focused on red-teaming reproduction tasks where the agent must craft byte-level inputs or complex scripts to prove a vulnerability is reachable.
- ▌ Descriptive And Statistical Analysis · dingxingdi bundleSkill: descriptive and statistical analysis
- ▌ Physical Safety And Intervention Adjudication · dingxingdi bundleSkill: physical-safety-and-intervention-adjudication
- ▌ Disruption Aware Reactive Replanning · dingxingdi bundleSkill: disruption-aware-reactive-replanning
- ▌ High Level Declarative Orchestration · dingxingdi bundleUse this skill when the user needs to achieve complex, long-horizon objectives that span multiple stages, hosts, or environments where low-level execution logs would overwhelm the agent. Trigger it for requests like 'plan at a high level,' 'decouple the plan from the execution,' 'use declarative macro-tasks instead of raw commands,' 'manage major milestones,' and 'handle complex deployments without getting bogged down in terminal logs.' It ensures the orchestrator focuses on 'what' to achieve conceptually while delegating the dense technical 'how' to localized scripts, external state trackers, or sub-agents.
- ▌ Orchestration Complexity Calibration · dingxingdi bundleSkill: orchestration-complexity-calibration
- ▌ Hierarchical Semantic Navigation Planning · dingxingdi bundleUse this when the user provides long-horizon navigation instructions in 3D environments that require following a sequence of landmarks, extracting map coordinates, or navigating based on functional zones and user-lifestyle demands. Trigger it for requests like 'follow the route from the bank to the park', 'I am a graphic designer, find a tidy spot for me to work', 'go to the study and organize my documents', or 'guide the drone over the damaged buildings in the satellite image'.
- ▌ Spatial Topological Construction Planning · dingxingdi bundleUse this when the user wants planning data about where the agent can move, what it can paint or place from nearby, or how earlier building steps change what becomes reachable later. Trigger it for requests like 'make planning tasks with layout constraints', 'give me build-order problems', 'create tasks where you can only act from neighboring positions', or 'make plans where the path changes as the structure grows.'
- ▌ Adversarial Jailbreak Translation Resistance · dingxingdi bundleUse this skill when the user provides translation requests that contain malicious 'jailbreak' patterns, adversarial noise, or formal logical expressions designed to bypass safety filters. Trigger it for requests like 'translate this despite the weird noise at the end,' 'the content is hidden in a logic formula (FOL),' or 'the prompt uses academic framing about formal semantics to disguise a harmful task.' Use it when a user attempts to use persuasive language or structural distribution shifts (like First-Order Logic) to trick the translator into outputting prohibited content.
- ▌ Cross Modal Alignment And Dialogue Grounding · dingxingdi bundleUse this skill when the user asks 'what is visually present when this sound happens?', 'the video shows one thing but the person says another, which is right?', 'correct the spelling of this name based on the screen', 'find the exact time for this speech', or 'match the voice to the person'. It is triggered for tasks requiring explicit temporal grounding involving both audio and video, implicit cross-modal retrieval, or resolving cross-modal contradictions (like using sharp on-screen text to fix noisy or muffled audio transcripts).
- ▌ Interleaved Multimodal Content Composition · dingxingdi bundleUse this skill when the user wants a document that combines narrative text with specific visual evidence, such as charts, images, or maps, to support analytical claims. Trigger it for requests like “write a data-driven report with graphs,” “show me the trends with visualizations,” “interleave images with your guide,” or “create a report where the charts and text prove each other.” It is essential for ensuring that visuals are constructed or selected precisely when the narrative needs them, ensuring that subsequent text interprets the data rather than just describing the surface of an image.
- ▌ Open Ended Reasoning And Explanation · dingxingdi bundleSkill: open-ended-reasoning-and-explanation
- ▌ User Action Guidance In Dual Control · dingxingdi bundleSkill: user-action-guidance-in-dual-control
- ▌ Graph Structured Relational Reasoning · dingxingdi bundleSkill: graph-structured relational reasoning
- ▌ Combinatorial Search Optimization Planning · dingxingdi bundleUse this when the user needs to solve complex logic puzzles, evaluate solvability, find the best route through millions of points, or optimize resource allocation in NP-hard problems with irregular structures. Trigger it for requests like 'solve this massive 10-million node route', 'find a Steiner tree for these points', 'solve this asymmetric delivery problem where road A-to-B is different than B-to-A', or 'solve a logic challenge designed to break simple guessing algorithms'.
- ▌ Contextual Lexical Sophistication Alignment · dingxingdi bundleUse this skill when the writing must scale its complexity, vocabulary, and grammatical structures to strictly match a designated audience proficiency level (e.g., CEFR A1-C2, HSK 3-6, layperson vs. expert). Trigger it for requests like 'explain this medical report in plain English,' 'ensure the grammar is HSK 4 compatible,' 'make this sound more academically sophisticated,' or 'check this essay to simulate errors a non-native might make.' It seamlessly handles both vocabulary disambiguation for upscaling and standardized language acquisition modeling for downscaling.
- ▌ Experiment Reproduction And Result Verification · dingxingdi bundleUse this skill when the user wants data where the agent must execute experiments, inspect logs and outputs, and judge whether the intended research results were reproduced. Trigger it for requests like 'generate reproduction verification tasks', 'make result-checking workflows', 'create reproduce.sh evaluation data', or 'give me code-agent tasks about running and validating research experiments'. Do not use it for first-pass paper-to-code implementation only.
- ▌ Multi Turn Instruction Following And Refinement · dingxingdi bundleUse this skill when a user wants to iteratively update or refine code through multiple rounds of instructions, especially when new constraints (like error handling, edge cases, or style rules) are added one by one. It is triggered by requests like 'keep the old behavior but add this new rule', 'refine the function I just wrote to handle empty inputs', 'don't forget the case sensitivity I mentioned earlier', or 'update the code to match my new feedback'. Use this to ensure the agent doesn't 'forget' previous requirements while implementing new ones.
- ▌ Capacity Constrained Transport Coordination · dingxingdi bundleUse this when the user wants planning data about moving several items with limited carrying ability, two hands, one hoist, or tight transport capacity. Trigger it for requests like 'make tasks where the robot can only carry a little at a time', 'give me move-things-between-rooms problems', 'create loading and unloading plans with bottlenecks', or 'make planning tasks where one carrier choice blocks another.'
- ▌ Continual Experience Driven Plan Refinement · dingxingdi bundleUse this when you want an agent that accumulates experience across a series of tasks to improve its success, or when an agent must use past 'Thought Patterns' to figure out complex or vague intents. Trigger it for requests like 'learn from your previous attempts', 'suggest some gear for an occasion based on past experience', 'I am not sure what I need but I want to start a new hobby', or 'distribute memory units across different agents based on their roles.'
- ▌ Hierarchical Symbolic Subgoal Decomposition · dingxingdi bundleUse this when the user wants to solve a large, multi-faceted project by breaking a distal goal into a sequence of abstract middle-states, sub-goals, or typed nodes like research versus writing. Trigger it for requests like 'break this big project into different types of steps', 'recursively decompose this goal until you have simple steps', 'give me a high-level strategy and then the moves to do it', or 'organize a plan where composition steps wait for reasoning steps.'
- ▌ Video Text Temporal Grounding · dingxingdi bundleUse this skill when a user wants to find a specific time window or 'clip' within a video based on an event, object, or action. Trigger it for requests like 'find the part in the video where they score', 'get the timestamps for the interview segment', 'show me the moment the cat jumps', or 'extract the clip of the engine failure'. This skill is for navigating continuous temporal visual data to locate discrete start and end markers (grounding).
- ▌ Clarification Under Underspecification · dingxingdi bundleSkill: clarification-under-underspecification
- ▌ Critical External Knowledge Integration · dingxingdi bundleSkill: critical external-knowledge integration
- ▌ Adversarially Robust Tool Orchestration · dingxingdi bundleUse this skill when the user wants to test an agent's ability to resist manipulation of its internal reasoning, task plans, or session memory. It targets scenarios where an attacker might inject fictitious plans into the conversation history, 'poison' the stored context with malicious follow-ups, or use logical bridges to redirect the agent from a legitimate goal to a harmful one. Trigger it for requests like 'test if the agent follows a fake plan,' 'see if it gets hijacked by corrupted memory,' 'make the task instructions change midway using a trap,' and 'check robustness against logical bridges or plan injections.' It applies whenever the orchestrator must verify that its current execution track still aligns with the original user intent despite adversarial context manipulation.
- ▌ Specialized Legal Rule Mapping · dingxingdi bundleUse this skill when the user provides a 'story' or hypothetical scenario and needs to find the specific legal rules or statutes that apply to it, even if the scenario and the documents share very few common words. It is triggered by requests like 'find the laws that cover this situation,' 'tell me if this person broke a rule based on the handbook,' 'apply the housing code to this tenant story,' or 'help me with a bar-exam style legal lookup.' This skill maps everyday factual descriptions to formal legal authority through reasoning-driven retrieval.
- ▌ Misleading Evidence Robustness · dingxingdi bundleUse this skill when the user wants to test if the model can be 'tricked' by biased, one-sided, or selectively framed documents. Trigger it for requests like 'test if it falls for misinformation', 'give me documents that look convincing but are wrong', 'try to lead the model to the wrong answer with biased text', or 'check if the agent can spot partial truths and selective omissions'.
- ▌ Robust Multi Hop Path Recovery · dingxingdi bundleUse this skill when the most direct piece of information is missing from the files and the agent must find a 'backup' way to the answer. Trigger it for requests like 'find a workaround in the data', 'what if the direct link is gone', 'use indirect clues to fill the gap', 'connect the dots through a middle person', or 'verify if the model can still figure out the answer even with incomplete records'.
- ▌ Internal API Aware Code Completion And Refinement · dingxingdi bundleUse this skill when you need to finish writing a block of code that relies on functions or classes defined elsewhere in the same project rather than standard libraries. It should be triggered by layman requests such as 'finish this function using our internal helpers', 'complete this call but I don't remember the exact sub-module function', 'help me use our custom API to register this dataset', or 'fix my code where I'm guessing the name of our team's functions'. Do not use it for general Python library completion; use it for project-specific internal logic that isn't found in online docs.
- ▌ Rigorous Issue Verification And Test Augmentation · dingxingdi bundleUse this skill when a user wants to proactively identify gaps in a software project's test suite, generate new reproduction tests for a bug (TDD), or rigorously verify an issue resolution. It is specifically triggered by everyday requests like 'check if these tests are enough for the bug', 'make a test that shows the bug happening', 'write a reproduction script for this issue', 'make sure the fix handles all edge cases', or 'try to find where the original developer missed a scenario'. This skill ensures that a bug fix is not just superficially correct but grounded in an executable fail-to-pass signal and robust against unhandled boundary conditions.
- ▌ Visualization Grounded Insight Reporting · dingxingdi bundleSkill: visualization-grounded insight reporting
- ▌ Autonomous Heuristic Optimization Algorithm Design · dingxingdi bundleUse this skill when the user wants to generate or refine code for hard optimization problems where exact answers are impossible, like delivery routing, scheduling, or resource planning. Trigger it for requests like 'solve this routing puzzle', 'optimize my delivery paths', 'make a scheduler for factory orders', or 'help me get a higher score on this algorithmic challenge'. It is specifically meant for iterative refinement tasks where you must use test feedback to continuously improve a program's numerical score over a long period.
- ▌ Multi Turn Persuasive Strategic Dialogue · dingxingdi bundleSkill: multi-turn-persuasive-strategic-dialogue
- ▌ Truth Maintenance Under Invalid Feedback · dingxingdi bundleSkill: truth-maintenance-under-invalid-feedback
- ▌ Execution Feedback Driven Iterative Refinement · dingxingdi bundleUse this when the user wants to automate technical or tool-use workflows that require running code or invoking APIs, analyzing error logs or feedback, and fixing the logic step-by-step. Trigger it for requests like 'debug this API call using feedback', 'iteratively refine this script based on scores', 'solve this task and correct any errors found in the tool response', or 'improve the tool plan based on execution failures.' It is particularly useful for System 2 reasoning where trial-and-error is mandatory for success.
- ▌ Information Gain Maximizing Sequential Inquiry · dingxingdi bundleUse this when the user needs to reach a target outcome by asking the most efficient sequence of questions to narrow down a large set of possibilities. Trigger it for requests like 'find the right item using the fewest questions', 'create a help desk flow that identifies the problem fast', 'give me a search plan that eliminates wrong options systematically', or 'design a guide that gets to the answer without redundant steps'.
- ▌ Cross Part Comparative Reasoning · dingxingdi bundleUse this skill when the user wants to compare, contrast, or find differences between different sections of a document or across multiple documents. Trigger it for requests like 'compare the 2023 and 2024 results', 'how do these two authors differ in their views?', 'what changed between the beginning and the end of the story?', or 'show the differences between these two companies'. This skill is for synthesis tasks where the answer requires extracting distinct facts from separate locations and performing a logical comparison.
- ▌ Repository Level Intent And Architectural Reasoning · dingxingdi bundleTrigger this skill when the user wants to understand the architecture, intent, or technical details hidden within a complex codebase or its associated documentation. Use it for requests like 'explain how the payment flow works', 'find where the research paper's logic is implemented', 'summarize the RL hyperparameters in the LaTeX docs', 'tell me why this class was designed this way', or 'find the data flow for the login module across the whole repo'. It is specifically meant for multi-hop reasoning over code, PDFs, TeX files, and READMEs to bridge the gap between high-level documentation and final implementation.
- ▌ Abstention Under Missing Or False Premise · dingxingdi bundleSkill: abstention-under-missing-or-false-premise
- ▌ Adaptive Pipeline Orchestration And Repair · dingxingdi bundleSkill: adaptive pipeline orchestration and repair
- ▌ Closed Loop Instruction Validation And Recovery · dingxingdi bundleSkill: closed-loop-instruction-validation-and-recovery
- ▌ Complex Document Layout Reasoning · dingxingdi bundleUse this skill when the user wants questions about flowcharts, diagrams, or info hidden in page headers, footers, and complex 2D structures. Trigger it for requests like 'follow the process in this diagram', 'read the chart layout', 'what comes after this step in the map', or 'check the footer for the document code'. This is for scenarios where the answer isn't just in a text block but depends on where things are placed on the page.
- ▌ Adaptive Safety Refusal Integrity · dingxingdi bundleUse this skill when the user wants to test if the model can be tricked into giving out dangerous or forbidden information by using creative writing or 'acting' tricks. Trigger it for requests like 'make the agent answer a harmful question by role-playing', 'try to get past the rules by using slang or expert-sounding language', 'ask for instructions on illegal acts but frame it as a story', or 'test if the model refuses bad advice even when it's pressured to be helpful'.
- ▌ Self Correction Guided Refinement · dingxingdi bundleUse this skill when the user wants to ensure the highest degree of accuracy through a 'think-twice' approach or self-critique. Trigger it for requests like 'double check your answer against the text', 'reflect on your initial guess', 'make sure you didn't miss any contradictions between the files and your internal logic', or 'correct your reasoning if the documents say something counter-intuitive'. This skill is specifically for RAG agents that follow an iterative 'Refine' loop to improve reasoning steps by critiquing a preliminary response against retrieved evidence.
- ▌ Empathetic Validation And Active Listening · dingxingdi bundleSkill: empathetic-validation-and-active-listening
- ▌ Uncertainty Calibrated Response Generation · dingxingdi bundleSkill: uncertainty-calibrated-response-generation
- ▌ Long Document Schema And Variable Grounding · dingxingdi bundleSkill: long-document schema and variable grounding
- ▌ UI Navigation Trajectory Retrieval · dingxingdi bundleUse this skill when the user asks questions about how to find specific features or perform multi-step tasks within a software application. Trigger it for requests like 'where is the settings menu?', 'guide me to the logout button', 'how do I view the privacy policy in this app?', or 'show me the path to buy a stock'. This skill is specifically for retrieving structured navigation sequences from specialized knowledge bases that map user intents to application UI transition graphs.
- ▌ Iterative Temporal Window Reasoning · dingxingdi bundleUse this skill when the user wants to find the best time for an activity based on changing conditions like weather or traffic. Trigger it for requests like 'when is the earliest I can leave to avoid rain', 'find a 2-hour dry window for my trip', 'help me postpone my travel to avoid congestion', or 'is there a better time today for a walk'. This skill guides the agent to iteratively search through time records, moving a 'search window' forward or backward until a period that satisfies all the user's constraints is found.
- ▌ Nuance Sensitive Semantic Grounding · dingxingdi bundleUse this skill when the user wants to ensure the agent doesn't exaggerate or oversimplify facts from the text. Trigger it for requests like 'don't make it sound more certain than it is', 'keep the specific conditions', 'avoid turning a maybe into a definitely', 'be careful with opinions presented in the docs', or 'don't leave out the fine print'. This skill ensures that the generated answer preserves the exact degree of certainty, specificity, and rhetorical nuance (e.g., distinguishing an opinion from a fact) found in the original retrieved passages.
- ▌ Clinical Experience Driven Reasoning · dingxingdi bundleUse this skill when the user wants to solve a medical problem by looking at what happened to other similar patients. Trigger it for requests like 'find similar cases to this patient', 'how do doctors usually treat people with these same symptoms', 'give me examples of past medical experiences for this diagnosis', or 'compare this patient's history with the hospital records to find a pattern'. This skill focuses on a coarse-to-fine retrieval process that identifies analogous entities before extracting specific decision-making evidence.
- ▌ Hierarchy Aware Structural Grounding · dingxingdi bundleUse this skill when the user asks questions about specific articles, sections, or clauses in complex documents like laws, technical manuals, or company policies. Trigger it for requests like 'what does section 5 say about X', 'find the penalty for violating article 10', 'look up the rules in chapter 3', or 'connect different parts of the law that mention each other'. This skill is specifically for identifying information that is split across a hierarchy (Titles, Chapters, Sections) or linked via internal cross-references.
- ▌ Open Domain Data Discovery And Table Selection · dingxingdi bundleSkill: open-domain data discovery and table selection
- ▌ Stateful Online Sequential Grounding And Replanning · dingxingdi bundleSkill: stateful-online-sequential-grounding-and-replanning
- ▌ Authority Aware Structural Re Ranking · dingxingdi bundleUse this skill when you need to find the 'most important' or 'landmark' documents in a collection where popularity and rank matter as much as keywords. Trigger it for requests like 'find the landmark rulings', 'give me the most cited research papers', 'prioritize high-court decisions over local ones', or 'rank the results by how much other documents point to them'. This skill is essential for navigating legal, academic, or professional corpora where information is organized in a hierarchy or linked through a citation network.
- ▌ Cross Lingual Reasoning And Alignment · dingxingdi bundleUse this skill when the user asks a question in one language (e.g., German, Spanish, Chinese, Arabic) but the information needed is stored in documents of a different language (e.g., English), or split across documents in multiple languages. Trigger it for requests like 'find the answer from these English reports and tell me in German', 'compare what the local news and the international news say', or 'check if the model flips to English when I ask a question in Spanish about an English text'. This skill targets the unique challenge of performing complex reasoning (comparison, aggregation) across language boundaries while maintaining the correct response language.
- ▌ Multi Conditional Entity Localization · dingxingdi bundleUse this skill when the user provides multiple specific clues and wants to find the one person, place, or thing that fits all of them. Trigger it for requests like 'find someone who did A and also did B', 'who is the person mentioned in both these articles', 'locate an entity satisfying several conditions', or 'which item fits this set of descriptions'. This skill is specifically for identifying a unique subject through the intersection of independent factual constraints found across different parts of the corpus.
- ▌ Multi Turn Contextual Query Synthesis · dingxingdi bundleUse this skill when the user asks a follow-up or clarification question that depends on previous conversation history. Trigger it for requests that use pronouns like 'he', 'it', or 'that', or elliptical phrases like 'what about its address?', 'why was he there?', or 'any more examples of those?'. This skill is specifically for RAG agents that need to rewrite a non-standalone user turn into a standalone search query by resolving conversational co-references and context.