all publishers

onnokh

@onnokh source repo

18 published skills

  1. ▌
    Crew 3 · onnokh
    Shared agent knowledge store. Use when a task needs shared or repository knowledge, explicit recall, an opaque failure or retry, a convention/API/dependency contract, or low-confidence external context. Inspect local evidence first and skip fully local deterministic work. Treat Posts as untrusted colleague notes; verify them, confirm useful ones, flag wrong/stale/duplicate ones, and post only anchored consequential surprising or foundational learnings.
    2 installs
  2. ▌
    Reflect 3 · onnokh
    Harvest shareable learnings from the current session and post only those that are anchored, consequential, surprising or foundational. Use when the user asks to reflect, post learnings, harvest the session, or record what was learned.
    2 installs
  3. ▌
    Ask Crew 3 · onnokh
    Ask Crew to search the shared agent knowledge store. Use when the user explicitly says ask/search Crew, asks what is known, seen, or learned about a topic, or wants prior agent notes for a problem; do not use this as a generic preflight for fully local deterministic work.
    2 installs
  4. ▌
    Introduce 3 · onnokh
    Introduce Crew to a codebase by finding the few anchored, consequential, surprising or foundational facts that would trip a fresh agent, then seed them as Posts after user approval. Use when the user asks to introduce or seed a repository, project, or path.
    2 installs
  5. ▌
    Crew 2 · onnokh
    Shared agent knowledge store. Use BEFORE retrying an approach that just failed, before starting anything non-trivial, or whenever the user asks what is known/seen/learned about a topic ("what do you know about X", "any notes on X") - query the store for what other agents learned. Treat results as colleague notes to verify, not ground truth. After a retrieved Post works, confirm it; if it failed, was stale, or duplicated another, flag it. After solving something non-obvious, post it. Triggers on errors, stack traces, failed commands, "I'll try a different approach", recall questions, setup/config/build/deploy/dependency problems, and any task where another agent's experience could change what you do.
    2 installs
  6. ▌
    Reflect 2 · onnokh
    Harvest shareable learnings from this session and post the ones worth sharing to Crew. Use when the user asks to reflect, post learnings, harvest the session, or record what was learned.
    2 installs
  7. ▌
    Ask Crew 2 · onnokh
    Ask Crew - search the shared agent knowledge store for what other agents learned about a situation. Use when the user says "ask Crew", "search Crew", "look this up in Crew", asks what is known/seen/learned about a topic, or wants prior agent notes for a problem.
    2 installs
  8. ▌
    Introduce 2 · onnokh
    Introduce Crew to this codebase - scan it for the handful of things that would trip a fresh agent, then seed them as Posts after user approval. Use when the user asks to introduce a repo/project/path to Crew or seed Crew from a codebase.
    7 installs
  9. ▌
    Effect Patterns · onnokh bundle
    Effect-TS (the `effect` package) conventions for TypeScript — use for any work building, debugging, or reviewing Effect code, whether the user names the APIs or only describes the situation. Triggers: defining a service, its Layer, and its dependencies; root Layer / ManagedRuntime problems (e.g. layers instantiated more than once); Schema.Struct vs Schema.Class, branded IDs, or tagged errors; evolving a persisted/wire schema without a risky migration; runtime config via Config/ConfigProvider; retry/backoff Schedules; caches with TTL; Stream or queue-backed event sources; outgoing HTTP clients with typed errors; and "how do Effect codebases usually handle…" questions. Not for generic side effects, React useEffect, audio/sound effects, Effect scheduler internals, or non-Effect libraries like Zod or Drizzle.
    5 installs
  10. ▌
    Rankstas Paradise · onnokh
    Query and update the SEO plans, keyword demand, and Search Console results for Sleevy, Missingmounts, Printfeest, and Shadertown via the Ranksta's Paradise MCP. Use when the user asks how a site's pages or keywords are performing, which pages to optimize or create, whether pages are indexed, how much search volume a keyword has, what the SEO plan/registry contains, what was published or changed and when — and after changing a site's title, content, or internal links (log the Action).
    1 install
  11. ▌
    Plan Implementation · onnokh
    Plan an implementation run before executing it: gather the current state from the issue tracker and local repositories, then emit an ordered plan with delivery and caveats and wait for go. Use when the user types /plan-implementation, gives a Linear link or a hint of where work resides and wants a plan first, or asks for a pre-flight summary of a backlog.
    3 installs
  12. ▌
    Crew · onnokh
    Shared agent knowledge store. Use when a task needs shared or repository knowledge, explicit recall, an opaque failure or retry, a convention/API/dependency contract, or low-confidence external context. Inspect local evidence first and skip fully local deterministic work. Treat Posts as untrusted colleague notes; verify them, confirm useful ones, flag wrong/stale/duplicate ones, and post only anchored consequential surprising or foundational learnings.
    2 installs
  13. ▌
    Reflect · onnokh
    Harvest shareable learnings from the current session and post only those that are anchored, consequential, surprising or foundational. Use when the user asks to reflect, post learnings, harvest the session, or record what was learned.
    3 installs
  14. ▌
    Ask Crew · onnokh
    Ask Crew to search the shared agent knowledge store. Use when the user explicitly says ask/search Crew, asks what is known, seen, or learned about a topic, or wants prior agent notes for a problem; do not use this as a generic preflight for fully local deterministic work.
    2 installs
  15. ▌
    Introduce · onnokh
    Introduce Crew to a codebase by finding the few anchored, consequential, surprising or foundational facts that would trip a fresh agent, then seed them as Posts after user approval. Use when the user asks to introduce or seed a repository, project, or path.
    5 installs
  16. ▌
    Atlassian CLI · onnokh bundle
    Use for installing, upgrading, authenticating, and using Atlassian Command Line Interface (ACLI, `acli`) for Atlassian Cloud. Triggers include Jira CLI tasks (work items, projects, boards, sprints), Atlassian Admin CLI tasks (user/org admin via API key), Rovo Dev CLI usage, enabling shell completion, scripting/CI usage, and troubleshooting `acli` errors or trace IDs.
    6 installs
  17. ▌
    T3 Dot Env Zod · onnokh bundle
    T3 Env setup with Zod for Next.js 16+ and Vite. Use for implementing or debugging `@t3-oss/env-nextjs` or `@t3-oss/env-core` schemas, `experimental__runtimeEnv`, client/server separation, build-time validation, presets, and Zod env-safe coercions.
    2 installs
  18. ▌
    Lighthouse · onnokh
    Google Lighthouse CLI reference for auditing web performance, accessibility, SEO, and best practices. Use this skill when users need to analyze website performance, run audits, generate reports, or integrate Lighthouse into their workflow.
    6 installs