all publishers

axross

@axross source repo

83 published skills

  1. ▌
    Zod Schema 3 · axross bundle
    A change touching Zod — `z.object`, `z.infer`, `safeParse`, `z.codec`, `z.coerce`, `z.brand`, `z.discriminatedUnion`, `.refine`, `.transform`, `zodResolver`, `z.toJSONSchema`, `ZodError`, a schema module, a validation boundary, or an unvalidated `fetch` or `JSON.parse` result. For which inputs are untrusted at all, use an application-security capability. Modelling a type schema with Zod 4, decoding and encoding across a boundary, and enforcing validation where untrusted data enters. Covers the one-parse boundary, schema-derived types, codecs, coercion traps, error shape, and what a passing parse still does not make safe.
    1 repo stars
  2. ▌
    Code Review 3 · axross bundle
    Reviewing a code change — a pull request, a branch or commit-range diff, or a post-implementation self-review of your own work before calling it done. Not for writing the change, only for judging one that already exists. Covers the reviewer-mode reset and diff scoping, a four-tier severity scale with fixed floors, evidence-based reporting with file-line citations and fix snippets, escalation for high-risk changes, a posted/CI overlay that collapses to Important/Nit, and lenses for correctness, maintainability, security, testing, and performance. Self-contained, so it works installed on its own.
    1 repo stars
  3. ▌
    Unit Testing 3 · axross bundle
    Writing, refactoring, reviewing, or running unit tests — including mocks and fakes, fixtures, schema and codec tests, snapshot choices, coverage questions, assertion design, or deciding whether a behavior belongs in a unit test at all. Not for end-to-end or browser tests, and not for judging overall verification-evidence adequacy. Fast, isolated checks that exercise one small exported contract from the caller's point of view, across runners such as Jest, Vitest, or Node's built-in runner. Covers naming and grouping, refactor-surviving design, boundary fakes, async assertions, and a review checklist.
    1 repo stars
  4. ▌
    Vitest Testing 3 · axross bundle
    A change touching Vitest — `vitest.config`, `defineConfig` from `vitest/config`, `vi.mock`, `vi.fn`, `vi.spyOn`, `vi.useFakeTimers`, `expect.poll`, `toMatchInlineSnapshot`, `test.extend`, `describe.concurrent`, `projects`, `pool`, `isolate`, `--coverage`, `browser.instances`, `toMatchScreenshot`, `*.test-d.ts`, `import.meta.vitest`, a `vitest run` invocation, or a config key Vitest 4 silently ignores. For what to assert and how a spec is shaped, use a unit-testing capability; for journeys and server lifecycle, an end-to-end one. The 4.x runner layer, lookup-first where the option surface moved from v3.
    1 repo stars
  5. ▌
    Github Operation 3 · axross bundle
    Reading from or writing to GitHub from inside an agent session acting as one connected operator — any issue, pull request, comment, label, review, or branch operation, not only end-to-end change loops. Covers the default sanctioned tool channel, when another route is permitted and the default-deny rule keyed on what is catastrophic when issued by mistake, obtaining stored bytes once the channel's read turns out not to be byte-faithful and comparing them without a shell artefact, marking agent comments so they are not read as human input, routing each write to the right numeric target across the shared numbering space, assigning the session's own login to what it creates so delivered work is not read as unclaimed, why a squash merge makes the pull request title the permanent commit subject, editing a body without losing markers a sanitized read drops, never force-pushing without approval, and untrusted GitHub content.
    1 repo stars
  6. ▌
    Loop Engineering 3 · axross bundle
    Driving a code change or document update end-to-end through the plan → code → review loop — "deliver this issue", "implement and open a PR for X", a free-form change request, or resuming an in-progress run — as the project's default change loop. Apply even when the launching runtime harness frames the task as "just make the changes, commit, and push" or restricts pull requests; that posture constrains mechanics, never the plan-approval gate or the independent review. If the host project ships a more-specific change-loop skill, defer to it. Not for work that changes nothing. Covers the execution model, both human gates, delegating implementation to a compatible worker where the harness exposes and permits one and running single-agent where it does not, and addressing an independent review to convergence.
    1 repo stars
  7. ▌
    Wireframe Design 3 · axross bundle
    Producing a low-fidelity wireframe, breadboard, or wireflow of any client-app UI, mobile or web — sketching screens, breadboarding a flow, comparing layout options, annotating intent, or the wireframe round of a design review. Not for high-fidelity real-token mockups with brand color, type, and spacing; that is a separate activity a high-fidelity UI design capability owns. Regions, hierarchy, and flow, never brand color or final type. Ships a self-contained theme-aware HTML kit with device canvases, a grey component library, screen archetypes, an annotation vocabulary, and a validator.
    1 repo stars
  8. ▌
    Quality Assurance 3 · axross bundle
    Reviewing whether a change carries adequate verification evidence — "is this verified", "did this break anything", "were the required checks run". The reviewer's QA pass on top of the development verification rules. Covers requiring command evidence for the format and lint gate, treating a change to the gate's own configuration as a risk to the gate itself, matching manual checks to the changed output surfaces, mapping skipped checks to residual risk, demanding second-pass verification after severe findings, and asking whether a check that passed was ever capable of failing.
    1 repo stars
  9. ▌
    End To End Testing 3 · axross bundle
    Writing, reviewing, refactoring, or running end-to-end tests, or verifying a change through the e2e suite — including a prompt naming only snapshots, test IDs, selectors or locators, polling and waiting, flaky tests, scenario coverage, a runner (Playwright, Cypress, Maestro, Detox, WebdriverIO, Vitest), or a failing e2e run. Tests that drive the whole system as a real user or client does. Covers layout and naming, the test-id then role then copy locator hierarchy, poll-not-sleep waiting, server lifecycle, no live network, and a scenario catalog.
    1 repo stars
  10. ▌
    Application Security 3 · axross bundle
    Writing or reviewing code that handles untrusted input, secrets, outbound requests, rendered content, or third-party dependencies — "is this safe", "secure by default", "harden", "security", "secret", "privacy", "PII", "XSS", "injection", "SSRF", "safe fetch", "access control", or a dependency review. The OWASP Top 10:2025 lens in two modes, writing secure-by-default code and judging the risk a change introduces. Covers secrets and environment variables, input validation, output encoding, SSRF, access control and data exposure, and supply-chain risk.
    1 repo stars
  11. ▌
    Code Maintainability 3 · axross bundle
    Writing, refactoring, or reviewing code for maintainability or design — "readable", "too long", "refactor", "abstraction", "cohesion", "magic number", "dead code", "what should this be called", or "should this live elsewhere". Cohesion as the test of what belongs in one unit, naming and file organization with a fallback identifier vocabulary, route-local versus shared abstraction boundaries, complexity and readability limits, self-explanatory implementation, magic-number and dead-code discipline, scope control, and SOLID/DRY/KISS/YAGNI judgment. While authoring, each concern is a practice to uphold; while reviewing, a finding to raise against the diff.
    1 repo stars
  12. ▌
    Conventional Commits 3 · axross bundle
    Authoring or revising a Git commit message, titling a pull request, or judging whether a header conforms — "write a commit message", "what type should this be", "is this header valid", "does this need a scope", "mark this as a breaking change", or preparing a squash-merge title that becomes a permanent commit subject. Not the pull request body, which a pull-request-description capability owns, and not issue titles. Conventional Commits v1.0.0 — the header contract, types, scopes, footers, breaking-change markers, SemVer correlation, and a bundled validator.
    1 repo stars
  13. ▌
    Next App Development 3 · axross bundle
    Writing or reviewing a Next.js App Router change — "app router", "server component", "use client", "use cache", "server action", "route handler", "proxy", "middleware", "generateMetadata", "revalidateTag", "next/image", "next.config", "instrumentation", "cacheComponents", a hydration or serialization error, or a route rendering dynamically when it should be static. For a component's own composition, props, and state use a React component development capability; for CSS and themes a React component styling one. The version 16 framework layer; Pages Router out of scope.
    1 repo stars
  14. ▌
    Software Development 3 · axross bundle
    Every task that touches a project — implementing, refactoring, running a project command, or writing a pull request body — even when the request never mentions formatting, linting, comments, dependencies, docs, or descriptions. Also when you need to know how to run one of the project's operations (tests, dev server, build, lint, deploy) and must find it documented or ask. Not for a session that touches nothing, where the conduct baseline applies instead. The baseline discipline — the format/lint loop, scoped change management, risk-surface mapping, and contributor docs.
    1 repo stars
  15. ▌
    Agent Skill Authoring 3 · axross bundle
    Creating, refining, splitting, consolidating, renaming, or auditing an agent skill — drafting a `SKILL.md`, editing frontmatter, tightening a `description`, deciding where a new rule belongs, or running the structure validators. Triggers on "add a skill", "split this skill", "audit skills", "recast this skill as a capability", and any change to a `SKILL.md` or its `references/`. The authoring rules for the agentskills.io format — capability framing, discovery metadata that survives a host's listing truncation, section anatomy, progressive disclosure, cross-references, and three bundled validators, one per kind of edit.
    1 repo stars
  16. ▌
    Professional Behavior 3 · axross bundle
    Apply this in every session — answering a question, investigating, reviewing, or building. Triggers on any uncertainty about facts, scope, or intent; on "are you sure", "don't guess", "what's the latest"; before asserting a version, API, price, or date; before putting a decision to the human; and whenever a result is reported. Not a change-loop skill — it governs conduct within work already underway. Covers the three-source triage (look it up, research it, ask), the clarifying interview, asking through the harness's question tool, researching over recalling, replying in the language of the human's own message, accuracy discipline, and reporting that leads with the answer.
    1 repo stars
  17. ▌
    Agent Skill Management 3 · axross bundle
    Deciding whether material belongs in a skill at all, or in the project's own documentation instead, before deciding its tier; adding, editing, renaming, moving, or removing an agent skill in a project that keeps skills in two tiers; a `git status` showing installed copies or `skills-lock.json` out of sync with their source; or a skill you loaded turning out to be wrong, outdated, or missing a rule — including mid-task, and including when its upstream is one you do not own. The skill-or-document question, the storage and install model, the drift check, and how to route a defect found in an installed skill.
    1 repo stars
  18. ▌
    Sentry Instrumentation 3 · axross bundle
    A change touching Sentry — `Sentry.init`, `@sentry/react`, `@sentry/nextjs`, `@sentry/react-native`, `withSentryConfig`, `getSentryExpoConfig`, a DSN, an auth token, source maps, `sendDefaultPii`, `tracesSampleRate`, `replaysOnErrorSampleRate`, `enableLogs`, `beforeSend`, `captureException`, `onRequestError`, a tunnel route, a release, or an unsymbolicated stack trace. For log levels, catch placement, and which failures are worth capturing at all, use a software instrumentation capability instead. The vendor layer — SDK choice, identity, source maps, data-collection posture, sampling, and per-platform init. Lookup-first on versions.
    1 repo stars
  19. ▌
    High Fidelity UI Design 3 · axross bundle
    Designing, building, or reviewing a high-fidelity user-facing surface with real colors, type, spacing, and states — "visual design", "design tokens", "dark mode", "theme", "contrast", "a11y", "typography", "spacing scale", "touch target", "focus state", "interaction states", "loading state", "make it production-ready". For low-fidelity regions-and-flow wireframes, use a wireframe or breadboard approach instead. Covers semantic tokens and dark mode, visual hierarchy, the 8px grid, type scales, WCAG contrast, complete state sets, focus, and motion.
    1 repo stars
  20. ▌
    React Component Styling 3 · axross bundle
    Writing, reviewing, or refactoring the styles of a React component on web or mobile native — a CSS Module, a Unistyles `StyleSheet.create`, a theme or token file, or global styles. Triggers on "styling", "CSS", "theme", "design token", "dark mode", "container query", "breakpoint", "responsive", "touch target", "hover on mobile", "safe area", "style prop", "className", "colour gamut", "P3", "@supports", "browser support", "Baseline", "fallback", "reduced motion", or a surface wrong at some width, pointer type, or colour scheme. For design rationale — hierarchy, contrast targets, motion taste — use a high-fidelity UI design capability instead.
    1 repo stars
  21. ▌
    Software Instrumentation 3 · axross bundle
    Writing, reviewing, or modifying code that logs, throws, catches, reports an error, tracks a product event, identifies a user, or configures a logger, error tracker, tracing, or analytics tool — including a request naming only "logging", "log level", "capture exception", "error boundary", "breadcrumb", "trace sampling", "metrics", "cardinality", "analytics", "event name", "event schema", "identify", "reset on logout", "cookie consent", or an unhandled exception. The three telemetry signals plus the error handling that makes them actionable, and product-event tracking end to end. Names roles rather than SDKs, so a vendor capability supplies the mechanics.
    1 repo stars
  22. ▌
    React Component Development 3 · axross bundle
    Building or reviewing a React component — "props", "spread props", "compound component", "variant prop", "empty state", "error state", "skeleton", "data-testid", "testID", "extract a hook", "useEffect", "where state lives", "memo", "virtualize", "FlatList", or a surface that re-renders or scrolls badly. For styling — CSS Modules, Unistyles, tokens — use a React component styling capability instead. Composition, props contracts, extracted logic, state, memoization, loading and error surfaces, testability, and list virtualization, on web and mobile native alike.
    1 repo stars
  23. ▌
    Living Project Documentation 3 · axross bundle
    Updating a project's own documentation when a change alters what it says, and reading it before planning one — a behaviour, a domain term, or a decision that constrains future work; also placing or updating a project's conventions or operational procedures once they share one docs/ tree with that documentation. Triggers on "docs/", "spec", "domain model", "glossary", "ubiquitous language", "ADR", "supersede", "the docs are stale", "is this documented anywhere", "conventions/", "operations/". Not spec-driven development — a plan drives the change and this records what became true, so "spec-first" and "generate from the spec" route elsewhere. Covers the docs/ shape, its invariants, decision records, conventions and operations, and five single-purpose validators.
    1 repo stars
  24. ▌
    Technical Document Authoring 3 · axross bundle
    Drafting, structuring, editing, or reviewing a technical document — a design doc, RFC, ADR, runbook, README, API reference, changelog, or post-mortem. Triggers on "make this clearer", "how should I structure this", "is this readable", "too wordy", "passive voice", "hard to follow", "tighten this prose". While authoring, each rule is a practice to uphold; while reviewing, a finding to raise against the draft. Covers naming the reader, the four Divio document types, leading with the answer, sentence-level style, choosing between prose and tables, and a plain-English vocabulary ceiling. Self-contained.
    1 repo stars
  25. ▌
    Product Requirement Document Authoring 3 · axross bundle
    Writing, refining, or reviewing a PRD, feature spec, plan document, RFC, design doc, or issue description — including the plan-writing or issue-drafting step of any delivery workflow. Triggers on "write a PRD", "draft a spec", "refine this issue", "structure this plan", "write acceptance criteria", "how do I verify this is done", "is this requirement testable". Not for writing the code, only the document that specifies it. A canonical tool-agnostic section structure and the craft behind each section.
    1 repo stars
  26. ▌
    Zod Schema 2 · axross bundle
    Modelling data with Zod and enforcing validation where untrusted data enters — the schema layer, from one parse at the boundary to the type every caller downstream reads. Triggers on `z.object`, `z.infer`, `safeParse`, `z.codec`, `z.coerce`, `z.brand`, `z.discriminatedUnion`, `.refine`, `.transform`, `zodResolver`, `z.toJSONSchema`, `ZodError`, a schema module, or an unvalidated `fetch` or `JSON.parse` result. For whether an input is untrusted at all, use an application-security capability; for where the parse hook lives, the framework's own. Covers codecs, coercion traps, and what a passing parse still does not make safe.
    1 repo stars
  27. ▌
    Code Review 2 · axross bundle
    Reviewing a code change — a pull request, a branch or commit-range diff, or a post-implementation self-review of your own work before calling it done. The methodology for judging whether a change already written is safe to merge, and for reporting why. Not for writing the change, only for judging one that already exists. Covers the reviewer-mode reset, diff scoping, a four-tier severity scale with fixed floors, file-line evidence with fix snippets, escalation for high-risk changes, and lenses for correctness, maintainability, security, testing, and performance. Self-contained, so it works installed on its own.
    1 repo stars
  28. ▌
    Jest Testing 2 · axross bundle
    Working on Jest itself — the runner layer rather than the tests' content, covering its config, API, CLI, and transform and isolation mechanics. Triggers on a Jest config, `jest.fn`, `jest.spyOn`, `jest.mock`, `unstable_mockModule`, `useFakeTimers`, `toMatchSnapshot`, `moduleNameMapper`, `transformIgnorePatterns`, `next/jest`, `jest-expo`, a `--runInBand` or `--detectOpenHandles` run, or `Cannot use import statement outside a module`. For what is worth testing and the unit-versus-e2e call, use a unit-testing capability; for browser or device suites, an end-to-end-testing one. Baseline Jest 30.4.2.
    1 repo stars
  29. ▌
    Unit Testing 2 · axross bundle
    Writing, refactoring, reviewing, or running unit tests — fast, isolated checks that exercise one small exported contract from the caller's point of view. Triggers on mocks and fakes, fixtures, schema and codec tests, snapshot choices, coverage questions, assertion design, spec naming, or whether a behavior belongs in a unit test at all. For runner configuration and mocking APIs, use that runner's own capability — Jest, Vitest; for browser or device journeys, an end-to-end-testing one; for verification-evidence adequacy, a quality-assurance one. Covers refactor-surviving design, boundary fakes, and async assertions.
    1 repo stars
  30. ▌
    Vitest Testing 2 · axross bundle
    Working on Vitest itself — the 4.x runner layer rather than the tests' content, covering config, the `vi` API, pools, coverage, and Browser Mode, plus driving it without hanging a session. Triggers on `vitest.config`, `vi.mock`, `vi.useFakeTimers`, `expect.poll`, `toMatchInlineSnapshot`, `test.extend`, `projects`, `pool`, `browser.instances`, `toMatchScreenshot`, `*.test-d.ts`, `import.meta.vitest`, or a config key Vitest 4 silently ignores. For what to assert and how a spec is shaped, use a unit-testing capability; for journeys and server lifecycle, an end-to-end one. Lookup-first where the option surface moved from v3.
    1 repo stars
  31. ▌
    Github Operation 2 · axross bundle
    Reading or writing GitHub from an agent session sharing one connected operator identity — any issue, pull request, comment, label, review, or branch operation. For plan storage, issue/PR association, state encoding, marker selection, evidence destinations, reviewer configuration, and draft/ready publication, use the project's delivery guidance; for state meaning and readiness, use its change-loop practices. Covers channel qualification, bounded authenticated fallback, operator/bot attribution, target identification, authorization, stored-body fidelity, read-back, partial and unknown outcomes, COMMENT reviews, and append-only history.
    1 repo stars
  32. ▌
    Loop Engineering 2 · axross bundle
    Driving a code change or document update through an approved plan → execution → verification → independent review loop, including resuming or recovering an in-progress run. Owns phase progression, approval revision, evidence, finding state, and recovery meaning—not host execution APIs or delivery storage. Project policy chooses gates and branch rules; host instructions and tool usage conditions always prevail. Not for read-only questions, investigations, or reviews that change nothing.
    1 repo stars
  33. ▌
    Wireframe Design 2 · axross bundle
    Producing a low-fidelity wireframe, breadboard, or wireflow of any client-app UI, mobile or web — regions, hierarchy, and flow settled before anything commits to brand color or final type. Triggers on sketching screens, breadboarding a flow, comparing layout options, annotating intent, or the wireframe round of a design review. Not for high-fidelity real-token mockups with brand color, type, and spacing; that is a separate activity a high-fidelity UI design capability owns. Ships a self-contained theme-aware HTML kit with device canvases, a grey component library, screen archetypes, an annotation vocabulary, and a validator.
    1 repo stars
  34. ▌
    Quality Assurance 2 · axross bundle
    Reviewing whether a change carries adequate verification evidence — "is this verified", "did this break anything", "were the required checks run". The reviewer's QA pass on top of the development verification rules, judging the evidence a change offers rather than re-deriving the rule behind it. Covers requiring command evidence for the format and lint gate, treating a change to the gate's own configuration as a risk to the gate itself, matching manual checks to the changed output surfaces, mapping a skipped check to residual risk, and asking whether a check that passed was ever capable of failing.
    1 repo stars
  35. ▌
    End To End Testing 2 · axross bundle
    Writing, reviewing, refactoring, or running end-to-end tests, or verifying a change through the e2e suite — whole-system checks that drive a running build the way a real user or client does. Triggers on snapshots, test IDs, selectors or locators, polling and waiting, flaky tests, scenario coverage, a runner (Playwright, Cypress, Maestro, Detox, WebdriverIO, Vitest), or a failing e2e run. For a test that mocks its dependencies to isolate one unit, use a unit-testing capability; for the runner's own config and reporters, that runner's own. Covers the test-id then role then copy locator hierarchy, poll-not-sleep waiting, and server lifecycle.
    1 repo stars
  36. ▌
    Application Security 2 · axross bundle
    Writing or reviewing code that handles untrusted input, secrets, outbound requests, rendered content, or third-party dependencies — the OWASP Top 10:2025 lens in two modes, writing secure-by-default code and judging the risk a change introduces. Triggers on "is this safe", "secure by default", "harden", "security", "secret", "privacy", "PII", "XSS", "injection", "SSRF", "safe fetch", "access control", or a dependency review. Covers secrets and environment variables, input validation, output encoding, SSRF, access control and data exposure, and supply-chain risk.
    1 repo stars
  37. ▌
    Code Maintainability 2 · axross bundle
    Writing, refactoring, or reviewing code for maintainability or design — keeping what you change readable, cohesive, and cheap to change next time. Triggers on "readable", "too long", "refactor", "abstraction", "cohesion", "magic number", "dead code", "what should this be called", or "should this live elsewhere". While authoring, each concern is a practice to uphold; while reviewing, a finding to raise against the diff. Covers cohesion as the test of what belongs in one unit, naming and file organization, abstraction boundaries, complexity limits, dead code, and SOLID/DRY/KISS/YAGNI judgment.
    1 repo stars
  38. ▌
    Conventional Commits 2 · axross bundle
    Authoring or revising a Git commit message, titling a pull request, or judging whether a header conforms — the Conventional Commits v1.0.0 header contract, plus a validator that catches a malformed header before it reaches your history. Triggers on "write a commit message", "what type should this be", "is this header valid", "does this need a scope", "mark this as a breaking change", or preparing a squash-merge title that becomes a permanent commit subject. Not the pull request body, which a pull-request-description capability owns, and not issue titles. Covers types, scopes, footers, breaking-change markers, and SemVer correlation.
    1 repo stars
  39. ▌
    Expo App Development 2 · axross bundle
    Working on an Expo app's framework layer — the shell around the components, covering routes, app config, config plugins, permissions, safe areas, assets, on-device storage, testing, SDK upgrades, and shipping a build or update. Triggers on "expo-router", "prebuild", "CNG", "EXPO_PUBLIC_", "splash screen", "deep link", "expo-updates", "runtime version", "EAS", "expo-doctor". For a component's composition, state, or styling, use the React component capabilities. Derives no rules for `+api.ts` API routes, web or server output, native module authoring, brownfield integration, DOM components, or App Clips.
    1 repo stars
  40. ▌
    Next App Development 2 · axross bundle
    Writing or reviewing a Next.js App Router change — the version 16 framework layer, covering routing, rendering, caching, mutations, and deployment. Triggers on "app router", "server component", "use client", "use cache", "server action", "route handler", "proxy.ts", "middleware", "generateMetadata", "revalidateTag", "next/image", "next.config", "instrumentation", "cacheComponents", an auth interrupt, a hydration or serialization error, or a route rendering dynamically when it should be static. For a component's own composition, props, and state use a React component development capability; for CSS and themes a React component styling one. Pages Router out of scope.
    1 repo stars
  41. ▌
    Software Development 2 · axross bundle
    Every task that touches a project — implementing, refactoring, running a project command, or writing a pull request body. The baseline discipline underneath the specialised skills — keep the change scoped, run the format and lint loop, map what the change puts at risk, and find out how the project is actually run. Applies even when the request never mentions formatting, linting, comments, dependencies, docs, or descriptions, and when you need one of the project's operations — tests, dev server, build, lint, deploy — and must find it documented or ask. Not for a session that touches nothing, where the conduct baseline applies instead.
    1 repo stars
  42. ▌
    Agent Skill Authoring 2 · axross bundle
    Creating, refining, splitting, consolidating, renaming, or auditing an agent skill — portable content boundaries and agentskills.io authoring. Triggers on `SKILL.md`, frontmatter, `description`, "where does this rule belong", "split this skill", "recast this skill as a capability", and "audit skill portability". Skill management owns installation and active loading; project policy and host guidance own gates and execution. Covers capability framing, judgment versus mechanism, standard metadata versus host extensions, conditional references, a portability checklist, and structural validators.
    1 repo stars
  43. ▌
    Professional Behavior 2 · axross bundle
    Handling what you do not know, and handing back what you found — the conduct baseline underneath whatever else a session is doing, applying to a question answered in one turn as fully as to a feature built over many. Triggers on any uncertainty about facts, scope, or intent; on "are you sure", "don't guess", "what's the latest"; before asserting a version, API, price, or date; before putting a decision to the human; and whenever a result is reported. Not a change-loop skill — it governs conduct within work already underway. Covers the three-source triage (look it up, research it, ask), the clarifying interview, accuracy discipline, and reporting that leads with the answer in the human's own language.
    1 repo stars
  44. ▌
    Agent Skill Management 2 · axross bundle
    Deciding whether material belongs in a skill or project documentation, choosing its tier, and managing skill installation and active loading. Triggers on adding, editing, renaming, moving, or removing a skill; installed-copy or lockfile drift; same-name collisions, missing discovery, stale loaded content, or a loaded skill with a defective rule. Metadata authoring belongs to skill-authoring practices; host-specific reload and source-inspection commands belong to project operations. Covers source ownership, targeted installation, drift checks, and separate discovery and active-loading evidence.
    1 repo stars
  45. ▌
    Sentry Instrumentation 2 · axross bundle
    Wiring or reviewing a Sentry integration — the vendor layer, covering SDK choice, init, identity, source maps, sampling, and data-collection posture for a project that already knows what it wants captured. Triggers on `Sentry.init`, `@sentry/nextjs`, `@sentry/react-native`, `withSentryConfig`, `getSentryExpoConfig`, a DSN, an auth token, `sendDefaultPii`, `tracesSampleRate`, `replaysOnErrorSampleRate`, `beforeSend`, `onRequestError`, a tunnel route, session replay, cron monitors, or an unsymbolicated stack trace. For log levels, catch placement, and which failures are worth capturing at all, use a software instrumentation capability. Lookup-first on versions.
    1 repo stars
  46. ▌
    High Fidelity UI Design 2 · axross bundle
    Designing, building, or reviewing a high-fidelity user-facing surface with real colors, type, spacing, and states — the visual design system behind a production-ready surface, and the rationale behind each decision in it. Triggers on "visual design", "design tokens", "dark mode", "theme", "contrast", "a11y", "typography", "spacing scale", "touch target", "focus state", "interaction states", "make it production-ready". For low-fidelity regions-and-flow wireframes, use a wireframe or breadboard approach; for the stylesheet that implements the design, a React component styling capability. Covers semantic tokens, visual hierarchy, the 8px grid, type scales, WCAG contrast, complete state sets, focus, and motion.
    1 repo stars
  47. ▌
    React Component Styling 2 · axross bundle
    Writing, reviewing, or refactoring the styles of a React component on web or mobile native — the implementation mechanics of tokens, responsiveness, adaptive states, and platform stylesheets. Triggers on a CSS Module, a Unistyles `StyleSheet.create`, a theme or token file, global styles, "dark mode", "container query", "breakpoint", "responsive", "touch target", "hover on mobile", "safe area", "className", "colour gamut", "P3", "@supports", "Baseline", "reduced motion", or a surface wrong at some width, pointer type, or colour scheme. For design rationale — hierarchy, contrast targets, motion taste — use a high-fidelity UI design capability instead.
    1 repo stars
  48. ▌
    Software Instrumentation 2 · axross bundle
    Deciding what a system should report about itself, independent of the SDK that carries it out — the vendor-neutral layer over logs, metrics, traces, error handling, and product events. Triggers on code that logs, throws, catches, reports an error, tracks an event, or identifies a user, and on "log level", "capture exception", "error boundary", "breadcrumb", "trace sampling", "cardinality", "event name", "identify", "reset on logout", or "cookie consent". For which package to install, which option controls collection, or source maps and build-time tokens, use the installed vendor's own capability. Covers try/catch placement, metric cardinality, and event naming through consent.
    1 repo stars
  49. ▌
    Amplitude Instrumentation 2 · axross bundle
    Wiring or reviewing an Amplitude integration — the vendor layer, covering SDK init, identity and sessions, autocapture, tracking plans, consent, replay, ingestion, and cost. Triggers on `@amplitude/analytics-browser`, `@amplitude/unified`, `amplitude.init`, `amplitude.track`, `setUserId`, `autocapture`, `defaultTracking`, `Ampli`, `insert_id`, `MTU`, `serverZone`, `identityStorage`, session replay `sampleRate`, or an Amplitude API key. For vendor-neutral event naming or what-to-track questions, use a software-instrumentation capability instead.
    1 repo stars
  50. ▌
    Tanstack Query Development 2 · axross bundle
    Writing or reviewing anything that reads or writes server state through TanStack Query — the v5 server-state layer, owning cache identity, refetch timing, invalidation, and how a failure surfaces. Triggers on "useQuery", "useMutation", "queryOptions", "queryKey", "staleTime", "invalidateQueries", "setQueryData", "optimistic update", "infinite scroll", "QueryClient", a stale list after a write, a colliding cache entry, or a refetch that fires too often or never. For a component's own composition, props, and state use a React component development capability; for routing and rendering the matching app-framework one.
    1 repo stars
  51. ▌
    React Component Development 2 · axross bundle
    Building or reviewing a React component — how one is put together, from composition and props contracts through extracted logic, state, memoization, and testability, on web and mobile native alike. Triggers on "props", "spread props", "compound component", "variant prop", "empty state", "error state", "skeleton", "data-testid", "testID", "extract a hook", "useEffect", "where state lives", "memo", "virtualize", "FlatList", or a surface that re-renders or scrolls badly. For styling — CSS Modules, Unistyles, tokens — use a React component styling capability; for server state, a TanStack Query one.
    1 repo stars
  52. ▌
    Living Project Documentation 2 · axross bundle
    Updating a project's own documentation when a change alters what it says, and reading it before planning one — the `docs/` tree recording what the project is and how it works now, and the mechanism that corrects it when a change makes it wrong. Triggers on "docs/", "spec", "domain model", "glossary", "ubiquitous language", "ADR", "supersede", "the docs are stale", "is this documented anywhere", "conventions/", "operations/". Not spec-driven development — a plan drives the change and this records what became true, so "spec-first" and "generate from the spec" route elsewhere; the plan document itself belongs to a PRD capability, and sentence-level craft to a technical-writing one. Covers decision records, conventions and operations, and five single-purpose validators.
    1 repo stars
  53. ▌
    Technical Document Authoring 2 · axross bundle
    Drafting, structuring, editing, or reviewing a technical document — a design doc, RFC, ADR, runbook, README, API reference, changelog, or post-mortem — and the sentence-level craft that makes any of them readable. Triggers on "make this clearer", "how should I structure this", "is this readable", "too wordy", "passive voice", "hard to follow", "tighten this prose". Not for non-software writing — marketing copy, blog posts, talks, fiction, academic papers outside software — nor for a spec's section skeleton, which a PRD capability owns. Covers naming the reader, the four Divio document types, leading with the answer, and a plain-English vocabulary ceiling. Self-contained.
    1 repo stars
  54. ▌
    Product Requirement Document Authoring 2 · axross bundle
    Writing, refining, or reviewing a PRD, feature spec, plan document, RFC, design doc, or issue description — the canonical tool-agnostic section structure such a document takes, and the craft behind each section. Includes the plan-writing or issue-drafting step of any delivery workflow. Triggers on "write a PRD", "draft a spec", "refine this issue", "structure this plan", "write acceptance criteria", "how do I verify this is done", "is this requirement testable". Not for writing the code, only the document that specifies it; UI markup, module placement, and test mechanics stay with the project's own conventions.
    1 repo stars
  55. ▌
    Zod Schema · axross bundle
    Modelling data with Zod and enforcing validation where untrusted data enters — the schema layer, from one parse at the boundary to the type every caller downstream reads. Triggers on `z.object`, `z.infer`, `safeParse`, `z.codec`, `z.coerce`, `z.brand`, `z.discriminatedUnion`, `.refine`, `.transform`, `zodResolver`, `z.toJSONSchema`, `ZodError`, a schema module, or an unvalidated `fetch` or `JSON.parse` result. For whether an input is untrusted at all, use an application-security capability; for where the parse hook lives, the framework's own. Covers codecs, coercion traps, and what a passing parse still does not make safe.
    1 repo stars
  56. ▌
    Code Review · axross bundle
    Reviewing a code change — a pull request, a branch or commit-range diff, or a post-implementation self-review of your own work before calling it done. The methodology for judging whether a change already written is safe to merge, and for reporting why. Not for writing the change, only for judging one that already exists. Covers the reviewer-mode reset, diff scoping, a four-tier severity scale with fixed floors, file-line evidence with fix snippets, escalation for high-risk changes, and lenses for correctness, maintainability, security, testing, and performance. Self-contained, so it works installed on its own.
    1 repo stars
  57. ▌
    Jest Testing · axross bundle
    Working on Jest itself — the runner layer rather than the tests' content, covering its config, API, CLI, and transform and isolation mechanics. Triggers on a Jest config, `jest.fn`, `jest.spyOn`, `jest.mock`, `unstable_mockModule`, `useFakeTimers`, `toMatchSnapshot`, `moduleNameMapper`, `transformIgnorePatterns`, `next/jest`, `jest-expo`, a `--runInBand` or `--detectOpenHandles` run, or `Cannot use import statement outside a module`. For what is worth testing and the unit-versus-e2e call, use a unit-testing capability; for browser or device suites, an end-to-end-testing one. Baseline Jest 30.4.2.
    1 repo stars
  58. ▌
    Unit Testing · axross bundle
    Writing, refactoring, reviewing, or running unit tests — fast, isolated checks that exercise one small exported contract from the caller's point of view. Triggers on mocks and fakes, fixtures, schema and codec tests, snapshot choices, coverage questions, assertion design, spec naming, or whether a behavior belongs in a unit test at all. For runner configuration and mocking APIs, use that runner's own capability — Jest, Vitest; for browser or device journeys, an end-to-end-testing one; for verification-evidence adequacy, a quality-assurance one. Covers refactor-surviving design, boundary fakes, and async assertions.
    1 repo stars
  59. ▌
    Vitest Testing · axross bundle
    Working on Vitest itself — the 4.x runner layer rather than the tests' content, covering config, the `vi` API, pools, coverage, and Browser Mode, plus driving it without hanging a session. Triggers on `vitest.config`, `vi.mock`, `vi.useFakeTimers`, `expect.poll`, `toMatchInlineSnapshot`, `test.extend`, `projects`, `pool`, `browser.instances`, `toMatchScreenshot`, `*.test-d.ts`, `import.meta.vitest`, or a config key Vitest 4 silently ignores. For what to assert and how a spec is shaped, use a unit-testing capability; for journeys and server lifecycle, an end-to-end one. Lookup-first where the option surface moved from v3.
    1 repo stars
  60. ▌
    Github Operation · axross bundle
    Reading or writing GitHub from an agent session sharing one connected operator identity — any issue, pull request, comment, label, review, or branch operation. For plan storage, issue/PR association, state encoding, marker selection, evidence destinations, reviewer configuration, and draft/ready publication, use the project's delivery guidance; for state meaning and readiness, use its change-loop practices. Covers channel qualification, bounded authenticated fallback, operator/bot attribution, target identification, authorization, stored-body fidelity, read-back, partial and unknown outcomes, COMMENT reviews, and append-only history.
    1 repo stars
  61. ▌
    Loop Engineering · axross bundle
    Driving a code change or document update through an approved plan → execution → verification → independent review loop, including resuming or recovering an in-progress run. Owns phase progression, approval revision, evidence, finding state, and recovery meaning—not host execution APIs or delivery storage. Project policy chooses gates and branch rules; host instructions and tool usage conditions always prevail. Not for read-only questions, investigations, or reviews that change nothing.
    1 repo stars
  62. ▌
    Wireframe Design · axross bundle
    Producing a low-fidelity wireframe, breadboard, or wireflow of any client-app UI, mobile or web — regions, hierarchy, and flow settled before anything commits to brand color or final type. Triggers on sketching screens, breadboarding a flow, comparing layout options, annotating intent, or the wireframe round of a design review. Not for high-fidelity real-token mockups with brand color, type, and spacing; that is a separate activity a high-fidelity UI design capability owns. Ships a self-contained theme-aware HTML kit with device canvases, a grey component library, screen archetypes, an annotation vocabulary, and a validator.
    1 repo stars
  63. ▌
    Quality Assurance · axross bundle
    Reviewing whether a change carries adequate verification evidence — "is this verified", "did this break anything", "were the required checks run". The reviewer's QA pass on top of the development verification rules, judging the evidence a change offers rather than re-deriving the rule behind it. Covers requiring command evidence for the format and lint gate, treating a change to the gate's own configuration as a risk to the gate itself, matching manual checks to the changed output surfaces, mapping a skipped check to residual risk, and asking whether a check that passed was ever capable of failing.
    1 repo stars
  64. ▌
    End To End Testing · axross bundle
    Writing, reviewing, refactoring, or running end-to-end tests, or verifying a change through the e2e suite — whole-system checks that drive a running build the way a real user or client does. Triggers on snapshots, test IDs, selectors or locators, polling and waiting, flaky tests, scenario coverage, a runner (Playwright, Cypress, Maestro, Detox, WebdriverIO, Vitest), or a failing e2e run. For a test that mocks its dependencies to isolate one unit, use a unit-testing capability; for the runner's own config and reporters, that runner's own. Covers the test-id then role then copy locator hierarchy, poll-not-sleep waiting, and server lifecycle.
    1 repo stars
  65. ▌
    Application Security · axross bundle
    Writing or reviewing code that handles untrusted input, secrets, outbound requests, rendered content, or third-party dependencies — the OWASP Top 10:2025 lens in two modes, writing secure-by-default code and judging the risk a change introduces. Triggers on "is this safe", "secure by default", "harden", "security", "secret", "privacy", "PII", "XSS", "injection", "SSRF", "safe fetch", "access control", or a dependency review. Covers secrets and environment variables, input validation, output encoding, SSRF, access control and data exposure, and supply-chain risk.
    1 repo stars
  66. ▌
    Code Maintainability · axross bundle
    Writing, refactoring, or reviewing code for maintainability or design — keeping what you change readable, cohesive, and cheap to change next time. Triggers on "readable", "too long", "refactor", "abstraction", "cohesion", "magic number", "dead code", "what should this be called", or "should this live elsewhere". While authoring, each concern is a practice to uphold; while reviewing, a finding to raise against the diff. Covers cohesion as the test of what belongs in one unit, naming and file organization, abstraction boundaries, complexity limits, dead code, and SOLID/DRY/KISS/YAGNI judgment.
    1 repo stars
  67. ▌
    Conventional Commits · axross bundle
    Authoring or revising a Git commit message, titling a pull request, or judging whether a header conforms — the Conventional Commits v1.0.0 header contract, plus a validator that catches a malformed header before it reaches your history. Triggers on "write a commit message", "what type should this be", "is this header valid", "does this need a scope", "mark this as a breaking change", or preparing a squash-merge title that becomes a permanent commit subject. Not the pull request body, which a pull-request-description capability owns, and not issue titles. Covers types, scopes, footers, breaking-change markers, and SemVer correlation.
    1 repo stars
  68. ▌
    Expo App Development · axross bundle
    Working on an Expo app's framework layer — the shell around the components, covering routes, app config, config plugins, permissions, safe areas, assets, on-device storage, testing, SDK upgrades, and shipping a build or update. Triggers on "expo-router", "prebuild", "CNG", "EXPO_PUBLIC_", "splash screen", "deep link", "expo-updates", "runtime version", "EAS", "expo-doctor". For a component's composition, state, or styling, use the React component capabilities. Derives no rules for `+api.ts` API routes, web or server output, native module authoring, brownfield integration, DOM components, or App Clips.
    1 repo stars
  69. ▌
    Next App Development · axross bundle
    Writing or reviewing a Next.js App Router change — the version 16 framework layer, covering routing, rendering, caching, mutations, and deployment. Triggers on "app router", "server component", "use client", "use cache", "server action", "route handler", "proxy.ts", "middleware", "generateMetadata", "revalidateTag", "next/image", "next.config", "instrumentation", "cacheComponents", an auth interrupt, a hydration or serialization error, or a route rendering dynamically when it should be static. For a component's own composition, props, and state use a React component development capability; for CSS and themes a React component styling one. Pages Router out of scope.
    1 repo stars
  70. ▌
    Software Development · axross bundle
    Every task that touches a project — implementing, refactoring, running a project command, or writing a pull request body. The baseline discipline underneath the specialised skills — keep the change scoped, run the format and lint loop, map what the change puts at risk, and find out how the project is actually run. Applies even when the request never mentions formatting, linting, comments, dependencies, docs, or descriptions, and when you need one of the project's operations — tests, dev server, build, lint, deploy — and must find it documented or ask. Not for a session that touches nothing, where the conduct baseline applies instead.
    1 repo stars
  71. ▌
    Agent Skill Authoring · axross bundle
    Creating, refining, splitting, consolidating, renaming, or auditing an agent skill — portable content boundaries and agentskills.io authoring. Triggers on `SKILL.md`, frontmatter, `description`, "where does this rule belong", "split this skill", "recast this skill as a capability", and "audit skill portability". Skill management owns installation and active loading; project policy and host guidance own gates and execution. Covers capability framing, judgment versus mechanism, standard metadata versus host extensions, conditional references, a portability checklist, and structural validators.
    1 repo stars
  72. ▌
    Professional Behavior · axross bundle
    Handling what you do not know, and handing back what you found — the conduct baseline underneath whatever else a session is doing, applying to a question answered in one turn as fully as to a feature built over many. Triggers on any uncertainty about facts, scope, or intent; on "are you sure", "don't guess", "what's the latest"; before asserting a version, API, price, or date; before putting a decision to the human; and whenever a result is reported. Not a change-loop skill — it governs conduct within work already underway. Covers the three-source triage (look it up, research it, ask), the clarifying interview, accuracy discipline, and reporting that leads with the answer in the human's own language.
    1 repo stars
  73. ▌
    Agent Skill Management · axross bundle
    Deciding whether material belongs in a skill or project documentation, choosing its tier, and managing skill installation and active loading. Triggers on adding, editing, renaming, moving, or removing a skill; installed-copy or lockfile drift; same-name collisions, missing discovery, stale loaded content, or a loaded skill with a defective rule. Metadata authoring belongs to skill-authoring practices; host-specific reload and source-inspection commands belong to project operations. Covers source ownership, targeted installation, drift checks, and separate discovery and active-loading evidence.
    1 repo stars
  74. ▌
    Sentry Instrumentation · axross bundle
    Wiring or reviewing a Sentry integration — the vendor layer, covering SDK choice, init, identity, source maps, sampling, and data-collection posture for a project that already knows what it wants captured. Triggers on `Sentry.init`, `@sentry/nextjs`, `@sentry/react-native`, `withSentryConfig`, `getSentryExpoConfig`, a DSN, an auth token, `sendDefaultPii`, `tracesSampleRate`, `replaysOnErrorSampleRate`, `beforeSend`, `onRequestError`, a tunnel route, session replay, cron monitors, or an unsymbolicated stack trace. For log levels, catch placement, and which failures are worth capturing at all, use a software instrumentation capability. Lookup-first on versions.
    1 repo stars
  75. ▌
    High Fidelity UI Design · axross bundle
    Designing, building, or reviewing a high-fidelity user-facing surface with real colors, type, spacing, and states — the visual design system behind a production-ready surface, and the rationale behind each decision in it. Triggers on "visual design", "design tokens", "dark mode", "theme", "contrast", "a11y", "typography", "spacing scale", "touch target", "focus state", "interaction states", "make it production-ready". For low-fidelity regions-and-flow wireframes, use a wireframe or breadboard approach; for the stylesheet that implements the design, a React component styling capability. Covers semantic tokens, visual hierarchy, the 8px grid, type scales, WCAG contrast, complete state sets, focus, and motion.
    1 repo stars
  76. ▌
    React Component Styling · axross bundle
    Writing, reviewing, or refactoring the styles of a React component on web or mobile native — the implementation mechanics of tokens, responsiveness, adaptive states, and platform stylesheets. Triggers on a CSS Module, a Unistyles `StyleSheet.create`, a theme or token file, global styles, "dark mode", "container query", "breakpoint", "responsive", "touch target", "hover on mobile", "safe area", "className", "colour gamut", "P3", "@supports", "Baseline", "reduced motion", or a surface wrong at some width, pointer type, or colour scheme. For design rationale — hierarchy, contrast targets, motion taste — use a high-fidelity UI design capability instead.
    1 repo stars
  77. ▌
    Software Instrumentation · axross bundle
    Deciding what a system should report about itself, independent of the SDK that carries it out — the vendor-neutral layer over logs, metrics, traces, error handling, and product events. Triggers on code that logs, throws, catches, reports an error, tracks an event, or identifies a user, and on "log level", "capture exception", "error boundary", "breadcrumb", "trace sampling", "cardinality", "event name", "identify", "reset on logout", or "cookie consent". For which package to install, which option controls collection, or source maps and build-time tokens, use the installed vendor's own capability. Covers try/catch placement, metric cardinality, and event naming through consent.
    1 repo stars
  78. ▌
    Amplitude Instrumentation · axross bundle
    Wiring or reviewing an Amplitude integration — the vendor layer, covering SDK init, identity and sessions, autocapture, tracking plans, consent, replay, ingestion, and cost. Triggers on `@amplitude/analytics-browser`, `@amplitude/unified`, `amplitude.init`, `amplitude.track`, `setUserId`, `autocapture`, `defaultTracking`, `Ampli`, `insert_id`, `MTU`, `serverZone`, `identityStorage`, session replay `sampleRate`, or an Amplitude API key. For vendor-neutral event naming or what-to-track questions, use a software-instrumentation capability instead.
    1 repo stars
  79. ▌
    Tanstack Query Development · axross bundle
    Writing or reviewing anything that reads or writes server state through TanStack Query — the v5 server-state layer, owning cache identity, refetch timing, invalidation, and how a failure surfaces. Triggers on "useQuery", "useMutation", "queryOptions", "queryKey", "staleTime", "invalidateQueries", "setQueryData", "optimistic update", "infinite scroll", "QueryClient", a stale list after a write, a colliding cache entry, or a refetch that fires too often or never. For a component's own composition, props, and state use a React component development capability; for routing and rendering the matching app-framework one.
    1 repo stars
  80. ▌
    React Component Development · axross bundle
    Building or reviewing a React component — how one is put together, from composition and props contracts through extracted logic, state, memoization, and testability, on web and mobile native alike. Triggers on "props", "spread props", "compound component", "variant prop", "empty state", "error state", "skeleton", "data-testid", "testID", "extract a hook", "useEffect", "where state lives", "memo", "virtualize", "FlatList", or a surface that re-renders or scrolls badly. For styling — CSS Modules, Unistyles, tokens — use a React component styling capability; for server state, a TanStack Query one.
    1 repo stars
  81. ▌
    Living Project Documentation · axross bundle
    Updating a project's own documentation when a change alters what it says, and reading it before planning one — the `docs/` tree recording what the project is and how it works now, and the mechanism that corrects it when a change makes it wrong. Triggers on "docs/", "spec", "domain model", "glossary", "ubiquitous language", "ADR", "supersede", "the docs are stale", "is this documented anywhere", "conventions/", "operations/". Not spec-driven development — a plan drives the change and this records what became true, so "spec-first" and "generate from the spec" route elsewhere; the plan document itself belongs to a PRD capability, and sentence-level craft to a technical-writing one. Covers decision records, conventions and operations, and five single-purpose validators.
    1 repo stars
  82. ▌
    Technical Document Authoring · axross bundle
    Drafting, structuring, editing, or reviewing a technical document — a design doc, RFC, ADR, runbook, README, API reference, changelog, or post-mortem — and the sentence-level craft that makes any of them readable. Triggers on "make this clearer", "how should I structure this", "is this readable", "too wordy", "passive voice", "hard to follow", "tighten this prose". Not for non-software writing — marketing copy, blog posts, talks, fiction, academic papers outside software — nor for a spec's section skeleton, which a PRD capability owns. Covers naming the reader, the four Divio document types, leading with the answer, and a plain-English vocabulary ceiling. Self-contained.
    1 repo stars
  83. ▌
    Product Requirement Document Authoring · axross bundle
    Writing, refining, or reviewing a PRD, feature spec, plan document, RFC, design doc, or issue description — the canonical tool-agnostic section structure such a document takes, and the craft behind each section. Includes the plan-writing or issue-drafting step of any delivery workflow. Triggers on "write a PRD", "draft a spec", "refine this issue", "structure this plan", "write acceptance criteria", "how do I verify this is done", "is this requirement testable". Not for writing the code, only the document that specifies it; UI markup, module placement, and test mechanics stay with the project's own conventions.
    1 repo stars