laststance
@laststance Organization
Public Skills
gif-analyzer
by laststance
Analyze animated GIF files by extracting and viewing frames as sequential video. Use when: - User mentions a GIF file path (e.g., "./demo.gif", "~/Downloads/animation.gif") - User wants to analyze or understand a GIF animation - User asks about motion, changes, or content in a GIF - User attaches or references a .gif file for analysis - User wants to examine a screen recording in GIF format - User invokes /gif slash command Keywords: "GIF", ".gif", "animation", "animated", "frames", "screen recording", "analyze gif", "gif analysis", "view gif", "gif content", "gif motion" Trigger patterns: - Natural language: "Analyze this GIF: ./demo.gif" - Slash command: /gif <path> or /gif <path> <message> When triggered, extract frames using the Python script, view frames in order, and interpret as continuous video sequence.
qa-team
by laststance
Launch comprehensive QA Agent Team for post-implementation verification. Tests 5 perspectives in parallel using AGENT_TEAMS: Visual Integrity, Functional Correctness (with Impact Propagation), Apple HIG Compliance, Edge Cases, and UX Sensibility. Falls back to Task tool if AGENT_TEAMS unavailable. Use when: - Implementation is complete and needs QA verification - User requests "run QA", "test everything", "quality check" - After /workflow completion for end-to-end validation - User says "check if the app works", "verify the UI" Keywords: "QA team", "quality assurance", "test everything", "visual check", "functional test", "edge cases", "UX review", "HIG compliance", "qa-team", "run QA", "verify app", "quality gate"
mentor
by laststance
Interactive code mentoring with pseudo-Plan mode architecture. AI deeply analyzes, designs, and presents a visual blueprint (diagrams + sequence flows). Human approves the plan, then writes all code with AI guidance. After "done", AI verifies behavioral correctness (not implementation conformity). Addresses "Comprehension Debt" - ensures developers understand every line they write. Primary use cases (existing codebase): - Modifying features in large codebases - Adding new functionality to existing projects - Fixing bugs with full context understanding - Refactoring with pattern comprehension Secondary use case: - Learning new tech stacks with hands-on coding Use when: modifying existing code, adding features, fixing bugs, refactoring, learning new framework, want guided hands-on coding, avoid the "wrote it but don't understand it" syndrome. Keywords: mentor, guided coding, modify, feature, bugfix, refactor, existing codebase, hands-on, comprehension, ownership, learning, plan mode
load
by laststance
Load project context from Serena MCP memory for session initialization. Discovers memories, reads project overview and critical rules, and validates context sufficiency. Portable across all Serena-enabled agents (Claude Code, Cursor, Windsurf, etc.). Use when: starting a session, resuming work, or needing project context. Keywords: session, load, initialize, context, memory, restore, resume
troubleshoot
by laststance
Diagnose and fix issues in code, builds, deployments, and system behavior. Always traces root cause, forms hypotheses before fixing, and validates after. Operates with introspection markers and hypothesis-driven debugging. Use when: - User reports a bug, error, or unexpected behavior - Build or compilation is failing - Tests are failing or flaky - Performance has degraded - Deployment or environment issues - User says "fix", "debug", "broken", "failing", "not working", "error" Keywords: troubleshoot, debug, fix, error, bug, failing, broken, crash, exception, not working
save
by laststance
Save session context to Serena MCP memory for cross-session persistence. Analyzes accomplishments, persists learnings as structured memories, creates checkpoint for future session restoration. Portable across all Serena-enabled agents (Claude Code, Cursor, Windsurf, etc.). Use when: ending a session, pausing work, saving progress, or preserving learnings. Keywords: session, save, persist, checkpoint, memory, store, context, end
explain
by laststance
Deep, systematic explanation of code, concepts, and system behavior. Always operates at advanced level with introspection markers and validation. Analyzes target thoroughly before explaining, verifies accuracy after. Use when: - User asks "what does this do", "how does this work", "explain this" - User wants to understand code, architecture, or a concept in depth - User points to a file, function, or system and asks for explanation - User says "explain", "walk me through", "teach me", "break down" Keywords: explain, understand, how, what, why, walk through, teach, break down, clarify
task
by laststance
Standard implementation workflow for all coding tasks. Executes a systematic 5-phase cycle: Investigate → Plan → Implement → Verify → Complete. Integrates Serena think checkpoints, introspection markers, and quality gates. Supports --frontend-verify flag for browser/app/CLI visual verification. Use when: - User asks to implement a feature, fix a bug, or refactor code - User provides a task that requires code changes - User says "do this", "build this", "fix this", "add this" - Any implementation work involving code editing Keywords: task, implement, build, fix, add, create, refactor, update, change
product-inspiration
by laststance
Provides UI/feature implementation inspiration with wireframes by researching top-tier apps. Implements ALL proposed patterns in _trials/ folder, then user picks the best one after hands-on evaluation. Use when user: - Asks for UI/feature implementation ideas or inspiration - Wants to know how top-tier apps implement a feature - Needs wireframe proposals for a feature - Requests inspiration, asks "how do other apps do this", "show patterns" 開発中のプロダクトの機能/UIインスピレーションを提供。 トップティアアプリを調査し、5パターン前後のデザイン提案を全て実装。 実物を見比べてから最良のパターンを選択する「Try All, Pick Best」ワークフロー。