1.2K skills · sorted by relevance
Review a Figma design, screenshot, or written description for usability, visual hierarchy, consistency, and accessibility. Receive specific findings, explanations, and prioritized improvements, with feedback tailored to the design stage and intended users.
Analyze meeting transcripts to understand speaking share, verbal habits, listening, and facilitation. Receive feedback grounded in specific quotes, practical suggestions for future meetings, and comparisons across periods when enough meeting data is available.
Review a FastMCP pull request for justified behavior, implementation correctness, and compatibility. Independently check the proposed change, assess automated checks and reviewer feedback, and report a verdict tied to the exact revision examined.
Run a structured sprint retrospective, a review of a short work cycle, using team feedback and delivery results. Identify shared themes and produce two or three prioritized improvements with owners, deadlines, and measures of success.
Monitor an open GitHub pull request over time, respond to review feedback, and address automated check failures. Apply and push fixes, then report whether it appears ready to merge or what still blocks progress.
Refine a working feature through browser feedback from the user. Collect observations through typed comments or live speech and drawing, apply focused usability fixes, and save the changes in local commits.
Review code changes or a pull request for bugs, effects on existing behavior, test gaps, and project standards. Receive findings supported by the code and a clear account of review coverage, with local fixes available when explicitly requested.
Group requests from Plain support threads, public reviews, and Slack into shared themes. Create or update one Linear issue per theme with representative quotes, mention counts, and sources, leaving priority and ownership decisions to people.
Disables PUA's always-on mode from Codex CLI and sets feedback frequency to zero. The alias updates the local configuration while preserving unknown fields, then reports completion with evidence from the executed command.
Enable PUA always-on mode from Codex CLI using a short command alias. It preserves existing configuration fields, restores feedback frequency when it is zero, and requires command output as evidence before reporting completion.
Investigate recurring problems by mapping feedback loops, the cycles through which effects influence their causes. Identify underlying structures, match known patterns, and assess where changes could help and what unintended effects they might produce.
Explore training workflows that reduce harmful model responses using a written set of principles. The skill covers self-critique, revised training examples, and RLAIF, reinforcement learning from AI feedback, with implementation examples and hardware guidance.
Ask external AI command-line tools to independently review a development phase plan. Collect feedback by reviewer in REVIEWS.md, then feed the results into GSD planning to support revisions.
Review a Figma design, screenshot, or written description for usability, visual hierarchy, consistency, and accessibility. Receive specific findings, explanations, and prioritized improvements, with feedback tailored to the design stage and intended users.
Get structured feedback on a legal memo, brief, paper, or exam essay without having it rewritten. Review organization, reasoning, clarity, and citation form, with three prioritized improvements and a record of recurring writing patterns.
Review legal essays using IRAC: issue, rule, application, and conclusion. Get feedback on missed issues, rule accuracy, reasoning, and organization, plus recurring patterns across practice sessions, without rewritten essays or model answers.
Build a custom fuzzer, a tool that tests code with many generated inputs, using LibAFL in Rust. Combine input mutation, execution feedback, scheduling, and target execution components when standard fuzzers offer insufficient control.
Review Canva slides, posters, and flyers for layout, wording, visual priorities, readability, and accessibility. Get concrete fixes ranked by importance and organized by page, while keeping the original design unchanged.
Organize unresolved feedback on GitHub pull requests, group changes by file or behavior, and apply selected fixes locally. Run relevant tests and summarize completed work and remaining issues, with replies and thread resolution requiring explicit user direction.
Design, implement, tune, and test action combat for web games with clear attack and defense timing. Define reliable hit outcomes, give defensive moves distinct purposes, and verify that visual feedback matches combat results in repeatable scenarios.
Design and implement sound feedback for Three.js or web games, from action cues to music transitions. Define which sounds take priority, respect volume controls, provide visual alternatives, and check browser behavior and audio performance.
Register AI agents on Starknet as NFTs, unique digital ownership records, using ERC-8004. Manage agent details, collect authorized feedback, and request or review third-party assessments through the standard’s identity, reputation, and validation registries.
Search YouTube by keyword and collect linked video and comment datasets through the BrowserAct API. Gather titles, views, comment text, likes, and reply counts to support audience feedback research and topic monitoring.
Collect Reddit posts and comments through the BrowserAct API using keywords, date filters, and sorting options. Results include votes, authors, timestamps, and reply relationships, providing structured source data for competitor research and reviews of community feedback.