992 个 · 按相关度排序
Write and debug code for further training large language models with TRL. Match training methods to dataset formats, configure supervised or preference-based training, and set up reward functions through Python or the command line.
中文简介正在准备中。当前显示英语简介。
Guides reinforcement learning, where models learn from rewards, using torchforge and PyTorch. Covers reward design, custom loss functions, and distributed training so researchers can experiment with algorithms while keeping infrastructure concerns separate.
中文简介正在准备中。当前显示英语简介。
Create and maintain code comments, API specifications, documentation sites, and user guides. Apply consistent formats, document parameters and errors, validate code examples, and report which parts of the code have been documented.
中文简介正在准备中。当前显示英语简介。
Write and debug code for further training large language models with TRL. Match training methods to dataset formats, configure supervised or preference-based training, and set up reward functions through Python or the command line.
中文简介正在准备中。当前显示英语简介。
Create reference documentation directly from source code for public functions, classes, command-line commands, and endpoints. Organize parameters, defaults, return values, errors, and runnable examples, then check that the documentation matches the implementation.
中文简介正在准备中。当前显示英语简介。
Audit PHP and other source code by listing calls to known dangerous functions, checking for enumeration gaps, and examining input paths and safeguards. Store risk assessments and verdicts in a searchable audit ledger.
中文简介正在准备中。当前显示英语简介。
Find where code lives, understand how it works, and trace connections between functions using a repository index. Results include file locations and line numbers to help plan edits and assess which other code a change could affect.
中文简介正在准备中。当前显示英语简介。
Find where code lives, understand how it works, and trace connections between functions using a repository index. Results include file locations and line numbers to help plan edits and assess which other code a change could affect.
中文简介正在准备中。当前显示英语简介。
Set up login, user management, and access rules in a Convex app. The skill connects the chosen authentication provider, links signed-in identities to app users when needed, and protects backend functions with identity and permission checks.
中文简介正在准备中。当前显示英语简介。
Add recurring jobs to a Convex app by scheduling internal functions in convex/crons.ts. The skill guides interval selection, keeps each run small and safe to repeat, and checks that the schedule appears in the dashboard.
中文简介正在准备中。当前显示英语简介。
Review code in a convex/ directory for security, access control, input and output validation, performance, and schema design. Receive findings grouped by severity, with explanations of each issue and suggested fixes before release.
中文简介正在准备中。当前显示英语简介。
Write and review Go tests using patterns suited to functions, file operations, and terminal interfaces. Cover multiple scenarios, check state changes, and compare rendered output with saved expected results while keeping slow integration tests separately skippable.
中文简介正在准备中。当前显示英语简介。
Build Native SDK app logic in TypeScript using explicit state, messages, and update functions. Explains the supported language subset, compiler diagnostics, module organization, and tests for logic that compiles directly into native code.
中文简介正在准备中。当前显示英语简介。
Build dashboard panels in plain TypeScript with automatic retries, saved panel settings, and clear error states. Keep panel-specific helper functions in the same file to make each component easier to understand and maintain.
中文简介正在准备中。当前显示英语简介。
Build sidebars and widgets that receive data updates from their caller. This skill uses a single shared module instance, separate creation and update functions, and stored references to page elements for partial updates.
中文简介正在准备中。当前显示英语简介。
Build dashboard panels in vanilla TypeScript with resizing, saved layout state, loading indicators, and retry controls. Examples also show how to wrap existing data-fetching functions with caching and temporary request pauses after repeated failures.
中文简介正在准备中。当前显示英语简介。
Investigate Python failures in a sandbox, an isolated execution environment, by capturing partial logs, testing functions separately, and checking output files one at a time. Use the findings to identify failing steps and rebuild a working script.
中文简介正在准备中。当前显示英语简介。
Build a graph of functions and calls across multiple programming languages to investigate entry points, paths to sensitive operations, and potential impact. Use the results to prioritize security reviews and connect external findings with source code.
中文简介正在准备中。当前显示英语简介。
Create Mermaid diagrams from code relationships parsed by Trailmark. Visualize function calls, class inheritance, module dependencies, class members, complexity hotspots, and paths from inputs to sensitive functions, with adjustable focus, depth, and layout.
中文简介正在准备中。当前显示英语简介。
Inspect DWARF, the debug information stored with compiled code, to find functions and types, check structural integrity, and clarify specification details. Also helps write parsing tools using libraries such as pyelftools, libdwarf, or gimli.
中文简介正在准备中。当前显示英语简介。
Assess AWS, Azure, and GCP environments by identifying identities, permissions, and paths to broader access or data exposure. Covers cloud storage, secrets, cross-account trust, and CI/CD, the automated systems used to build and deploy software.
中文简介正在准备中。当前显示英语简介。
Query STRING for protein interaction networks, interaction partners, and enrichment results showing functions common in a gene set. Receive concise summaries with limited result counts, with options for machine-readable JSON or saved full results.
中文简介正在准备中。当前显示英语简介。
Test whether an AI agent reveals another user's data or performs actions beyond the caller's permissions. Dialogue probes cover access across users or organizations, admin functions, and attempts to gain higher privileges without authorization.
中文简介正在准备中。当前显示英语简介。
Review Apache Pinot changes for correctness in null handling, numeric types, type-based processing, and resource lifecycles. Use it to examine changes to calculations, aggregations, window functions, and segment cleanup or persistence.
中文简介正在准备中。当前显示英语简介。