alexleekt
- 16 skills
- 0 followers
- 15 hours ago last updated
- ▌ My Krea · alexleekt**ALWAYS use when user asks for:** image generation, video generation, creating pictures, making visual content, "generate image", "make a video", "create a picture", "AI image", "text to image", "text to video". **DO NOT use for:** photo editing or retouching, production-grade graphic design, video editing, or any non-generative visual tasks. Generate images and videos with Krea AI using the official Krea CLI.
- ▌ My Council · alexleekt**ALWAYS use when user mentions:** "review council", "multi-agent review", "parallel review", "team review", "full code review", "review with experts", "spawn reviewers", "council pattern", "review all dimensions", "code review", "review this", "PR review", "check my code", "critique", "how should I write this", "raise the bar". **ALWAYS use when:** a codebase has changed >5 files or ~100+ lines, or when the user wants a thorough review across multiple dimensions (security, performance, correctness, maintainability, tests, accessibility, architecture, types, research). **DO NOT use for:** single-file quick checks, asking how a specific function works, or debugging a known bug (use the appropriate debug skill instead).
- ▌ My Crawl4ai · alexleekt bundle**ALWAYS use when user mentions:** "crawl4ai", "web scraping", "crawl a site", "extract web content", "scrape data from URL", or needs to extract web content as markdown/JSON. **DO NOT use for:** Browser automation (clicking, filling forms, screenshots, testing web apps) — use @skills/agent-browser for those. Set up and use crawl4ai for web crawling/scraping using uv, just, and project-based workflows. Assumes uv for Python and just for task automation.
- ▌ My Workflow · alexleekt**ALWAYS use when:** assessing task direction, deciding whether to stay in the current worktrunk or switch; updating worktree names to match evolving work; managing session boundaries and context switching; spawning parallel subagents. Use whenever the user asks about worktrees, switching contexts, session starts/ends, or coordinating multiple agents. **DO NOT use for:** commit/push/ship actions — use @skills/my-vcs-hygiene. Plan/build/review/document phases — use @skills/my-project-lifecycle. `wt` CLI mechanics — use @skills/worktrunk.
- ▌ My Tech Stack · alexleekt**MANDATORY REFERENCE** for all tool recommendations. ALWAYS consult BEFORE suggesting tools, libraries, frameworks, or technologies. **ALWAYS use when user asks about:** preferred tools, tech recommendations ("bun or node", "uv or pip"), or stack decisions. **ALWAYS use when agent decides:** to install a package, add a dependency, choose a library/framework, select a build tool, or make any technology choice that affects the project's stack. **DO NOT use for:** agent configuration (@skills/my-agent-file-conventions) or VCS operations. **CRITICAL:** Do NOT suggest unlisted tools without asking first.
- ▌ My Vcs Hygiene · alexleekt**ALWAYS use when:** the user says commit, push, stage, ship, "looking good", "done", "wrap it up", or signals completion; when making commits; when starting a session with uncommitted changes or on a dirty main branch; when the user gives visual or UX feedback during building ("move this left", "make it smaller", "change the color"). Use whenever the user mentions committing, pushing, shipping, or wrapping up work — even if they don't explicitly use git terminology. Also use when deciding commit scope, handling monorepo cross-package changes, managing visual iteration loops, or amending commits. **DO NOT use for:** worktree switching or naming — use @skills/my-workflow. Plan/build/review phases — use @skills/my-project-lifecycle. `wt` CLI mechanics — use @skills/worktrunk.
- ▌ My Decision Log · alexleekt**ALWAYS use when:** a significant decision is made during a session (tool choice, architecture direction, library selection, workflow change, or approach pivot). Use when the user asks "why did we do X?", "document this decision", "what did we decide about Y?", or when creating ADRs. Also use proactively when a decision has trade-offs, is hard to reverse, or will affect future work. Track decisions with rationale so future sessions (or other agents) can understand the reasoning without re-deriving it. **DO NOT use for:** trivial decisions (variable names, formatting). Committed decisions with no trade-offs. Immediate technical fixes with obvious solutions.
- ▌ Cua Computer Use · alexleekt**ALWAYS use when user asks for:** computer-use, screen control, UI automation, app automation, clicking elements, typing text, taking screenshots, or controlling macOS apps programmatically. **Assumption:** Any request involving GUI interaction, screenshots, or app control on macOS should use cua-driver tools via this skill. **DO NOT use for:** headless/scripted automation that doesn't need visual/GUI interaction (use shell scripts instead).
- ▌ My Error Recovery · alexleekt**ALWAYS use when:** a tool fails, a command errors, a request times out, a file read returns nothing unexpected, or something goes wrong during execution. Use when the user says "it failed", "error", "something went wrong", "that didn't work", or when any tool call returns an error, stderr, or non-zero exit code. Also use when a skill produces unexpected output, a provider is unavailable, or the agent needs to pivot strategy after a failure. **DO NOT use for:** code review of error-prone code — use @skills/my-code-review. Preventing errors through planning — use @skills/my-project-lifecycle.
- ▌ My Web Search Kagi · alexleekt bundle**ALWAYS use when user asks for:** searching the web, looking up information online, finding docs, checking prices, researching a topic, verifying facts. Also use when no dedicated web search MCP server is available. Triggers on phrases like "search the web", "look up", "find online", "google", "kagi search", "web search", "search for", or when current/real-time information is needed. **DO NOT use for:** questions answerable from training data alone, tasks where the user explicitly says "don't search", or when a dedicated MCP search server is available and preferred.
- ▌ My Context Strategy · alexleekt**ALWAYS use when:** processing large outputs (logs, diffs, build output, API responses), analyzing files over 50KB, running commands that produce more than 20 lines, or deciding which tool to use for data gathering. Use when the user says "analyze logs", "process this output", "check build output", "count lines", "find TODOs", or any task involving filtering, counting, aggregating, or transforming data. Also use when deciding between bash, read, ctx_execute, ctx_execute_file, or ctx_batch_execute for a task. **DO NOT use for:** editing files — use Edit/Write directly. Navigating directories — use Bash cd/ls. Short fixed-output observation — use Bash directly.
- ▌ My Semantic Release · alexleekt**ALWAYS use when user mentions:** "release", "ship it", "cut a release", "changelog", "version bump", "semver", "conventional commit". **DO NOT use for:** daily VCS operations — use git. Automates semantic versioning, changelog generation, and release workflows.
- ▌ My Project Lifecycle · alexleekt**ALWAYS use when:** planning a new feature or system, deciding when to code review, documenting what was built, reflecting on a completed session to improve skills, or mapping the overall flow from plan → build → review → document → ship. Use when the user asks "how should we build this?", "when should we review?", "document what we did", "reflect on the session", or "update my skills". Also use when a feature feels complete and the user needs guidance on what comes next. **DO NOT use for:** committing or pushing — use @skills/my-vcs-hygiene. Worktree management — use @skills/my-workflow.
- ▌ My Communication Style · alexleekt**ALWAYS use when:** the user says "be more concise", "show me progress", "don't ask so much", "just do it", "too verbose", or gives any feedback about how the agent communicates. Also use when deciding whether to ask the user or act directly, how much detail to include in responses, when to show intermediate steps vs just results, or how to format output (tables vs lists vs prose). Use proactively when the user seems frustrated by verbosity, indecision, or lack of visibility. **DO NOT use for:** technical content decisions — use @skills/my-tech-stack. Workflow decisions — use @skills/my-workflow.
- ▌ My Session Retrospective · alexleekt**ALWAYS use when:** the user says "review this session", "how did we do", "assess adherence", "did we follow the workflow", or signals session completion. Also use when the user asks to evaluate skill invocation, check what skills were missed, or improve skill automatic triggering. Use at the end of any significant session (10+ turns, file mutations, or multi-phase work) to assess adherence against my-workflow, my-vcs-hygiene, my-project-lifecycle, and all other my-* skills. **DO NOT use for:** planning the next session — use @skills/my-project-lifecycle. Code review of output — use @skills/my-code-review.
- ▌ My Agent File Conventions · alexleekt**ALWAYS use when:** creating, editing, or reviewing AGENT.md or claude.md files. **ALWAYS use when:** deciding what belongs in agent configuration vs project docs. **ALWAYS use when:** the user says "agent file", "agent config", "behavioral rules", "claude.md", "AGENT.md", "what goes in agent files", or "agent guidelines". **DO NOT use for:** tech stack recommendations (@skills/my-tech-stack) or VCS workflows. Defines the boundary: behavioral rules (agent files) vs documentation (README, ARCHITECTURE). Covers file locations, templates, common mistakes, and the "does this belong?" decision tree.