All authors
Claude Skills by bakw00ds
github.com/bakw00ds67 skills0 installs18 views
- MyskillA test skill for parity testing.Votes: 0GitHub stars: 4
- A11y ScanRun axe-core / Pa11y / Lighthouse against changed pages and classify findings by WCAG level (A / AA / AAA). Use when shipping or reviewing UI changes, before a release accessibility gate, or when asked for an a11y audit.Votes: 0GitHub stars: 4
- About Page ScaffoldDrop in a 3-paragraph "what is this?" page for new users, per frontend framework. Use when a project lacks an onboarding/about page and you want to orient first-time users.Votes: 0GitHub stars: 4
- Adr WriteScaffold a new Architecture Decision Record in Michael Nygard format (Context / Decision / Consequences) and update the ADR index. Use when a non-trivial or hard-to-reverse architectural choice is made and needs a durable record.Votes: 0GitHub stars: 4
- Agent AuditScan dispatch-log for misuse patterns — lead-did-specialist-work, wrong-runtime, budget-violations, repeated-failures. Use when reviewing how a session dispatched work, or auditing for dispatch-discipline drift.Votes: 0GitHub stars: 4
- Api DiffDiff an OpenAPI spec across commits and classify each change as major/minor/patch per SemVer-for-APIs. Use before bumping an API version or releasing, or when checking whether a spec edit is breaking.Votes: 0GitHub stars: 4
- Architecture Viz ScaffoldGenerate a static architecture page that pulls ADRs, Mermaid diagrams, CHANGELOG, SBOM, and tech-debt log into one navigable UI, per frontend stack. Use when a project needs a single browsable view of its architecture artifacts.Votes: 0GitHub stars: 4
- Browser Testing With DevtoolsVerify front-end behavior in a live browser — DOM, console, network, computed styles, performance, accessibility tree — via the chrome-devtools MCP server, instead of trusting static code review or unit tests alone. Use when shipping a UI change, diagnosing a rendering/network/perf bug, or confirming a fix in the running app; requires the chrome-devtools MCP server to be configured.Votes: 0GitHub stars: 4
- Changelog Ui ScaffoldDrop in a UI component / page that surfaces the project's CHANGELOG.md and current version, per frontend framework. Use when you want users to see release history and the running version in-app.Votes: 0GitHub stars: 4
- Code SimplificationSimplify code for clarity while preserving exact behavior, understanding why code exists (Chesterton's Fence) before changing it. Use when recently-changed code is hard to read — deep nesting, long functions, nested ternaries, generic names, duplicated logic — and you want a behavior-preserving cleanup, NOT a feature change or a dead-code sweep.Votes: 0GitHub stars: 4
- Contract HandoffPublish a contract (API, DB schema, UI) for downstream teammates to consume. Use when one specialist's output is another's input and the interface must be fixed before dependent work starts.Votes: 0GitHub stars: 4
- Cost SummaryRoll up the yakos dispatch-log into a per-runtime / per-agent / per-day cost summary, optionally posting to a webhook. Use when reviewing spend, before a budget check, or when asked "what did this cost?".Votes: 0GitHub stars: 4
- Cve TriagePull CVEs against the current dependency set (osv.dev / GHSA) and classify each as exploitable / theoretical / not-applicable. Use during a security review, before release, or when a new advisory lands on a dependency.Votes: 0GitHub stars: 4
- Dependency UpdateSurvey and apply dependency updates safely, verifying each before adopting it. Use when dependencies are stale, a security update is needed, or during scheduled maintenance.Votes: 0GitHub stars: 4
- Deploy CheckPre-deploy verification — build, smoke test, env-var sanity. Use immediately before deploying or promoting to prod, to catch a broken build or missing config before it ships.Votes: 0GitHub stars: 4
- Deprecation And MigrationManage the sunset of an old system, API, or feature — decide maintain-vs-sunset, build a proven replacement, migrate consumers incrementally, and remove only after verifying zero usage. Use when proposing to deprecate or remove anything with consumers, when an old code path has become a maintenance liability, or when planning a migration off a legacy interface.Votes: 0GitHub stars: 4
- Design Tokens AuditScan the codebase for hardcoded values that should be design tokens (colors, spacing, font-size, radius) and surface drift against the registry. Use when reviewing UI styling or enforcing a design system.Votes: 0GitHub stars: 4
- Dispatch As Project AgentSpawn a generic Agent wearing a project-specific agent's discipline. Use when you need to dispatch a project-level agent that the v0.1 runtime can't natively resolve as a subagent_type.Votes: 0GitHub stars: 4
- Doubt Driven DevelopmentSubject a high-stakes or irreversible decision to a fresh-context adversarial review BEFORE it stands, by dispatching an independent reviewer who only sees the artifact and contract — not your reasoning. Use when a decision crosses module boundaries, asserts an unverifiable property (thread-safety, ordering, idempotence), or has irreversible blast radius (prod deploy, data migration, public API change).Votes: 0GitHub stars: 4
- Evidence Based DebuggingConstrain debugging to cite runtime evidence (logs, stack traces, repro output) before proposing fixes. Use when investigating a bug, an outage, or unexpected behavior — observe before you theorize, never guess-and-patch.Votes: 0GitHub stars: 4
- Feedback ScaffoldDrop in the full user-feedback subsystem (DB migration, API handlers, UI submit widget, view-my-feedback page, CHANGELOG citation wiring) per stack. Use when a project needs to collect and surface user feedback end-to-end.Votes: 0GitHub stars: 4
- Finops ReviewAnalyze the dispatch-log for per-feature spend, cache hit rate, and model routing, surfacing optimization opportunities. Use when costs look high, before a budget review, or when hunting for ways to cut LLM spend.Votes: 0GitHub stars: 4
- Flake QuarantineAuto-tag tests that flaked >N times in M runs, move them to a quarantine suite, and cap how long they stay there. Use when flaky tests are destabilizing CI and you need to isolate them without deleting coverage.Votes: 0GitHub stars: 4
- Gather FeedbackPull and triage user feedback from configured sources into actionable buckets. Use when prioritizing work from user input, before a planning cycle, or when asked "what are users saying?".Votes: 0GitHub stars: 4
- Hallucination CheckCitation-grounding verifier — for each claim in an LLM response, confirm support in the retrieved context and report ungrounded claims. Use when validating a RAG/LLM answer for factual grounding before trusting or shipping it.Votes: 0GitHub stars: 4
- Hashed EditHash-anchored line edits — refuse the edit if the target line content changed since the read. Use when editing a file that may have shifted between read and write, to avoid stale-line corruption.Votes: 0GitHub stars: 4
- Hook Bypass ReviewAudit work/current/hook-bypass.md for stale entries, expired bypasses, and patterns suggesting a hook needs tuning. Use before archive, periodically during long projects, or when hook bypasses are accumulating.Votes: 0GitHub stars: 4
- I18n AuditScan for hardcoded UI strings, missing translation keys, RTL layout breakages, untested locales, and date/number/currency hardcoding. Use when localizing a UI, adding a locale, or auditing internationalization readiness.Votes: 0GitHub stars: 4
- Interaction PatternsHeuristic eval against Nielsen's 10 usability heuristics plus a WCAG 2.2 first-pass for keyboard, focus, contrast, and motion. Use when reviewing a UI's interaction design and baseline accessibility together.Votes: 0GitHub stars: 4
- Interview MeClose the gap between what was asked and what is actually needed via one-question-at-a-time interviewing, each question carrying your guessed answer, until intent is ~95% clear. Use when a request is underspecified on who/why/success-criteria/constraints, when you're tempted to fill in unstated assumptions, or when the user says "interview me" / "stress-test my thinking" before any building begins.Votes: 0GitHub stars: 4
- Iterate UntilLoop a work-then-verify cycle until a human-checkable verifier passes, with a hard iteration cap and an audit trail. Use when a task has a concrete pass/fail check and should be retried automatically until it succeeds.Votes: 0GitHub stars: 4
- Kanban TendMaintain the project kanban — add tasks at dispatch, move on lifecycle events, archive completed work at release time. Use when dispatching work, on a blocker/completion, or when the board needs tidying.Votes: 0GitHub stars: 4
- License AuditScan the dependency tree for license-policy violations — copyleft in proprietary, unknown licenses, license downgrades. Use before release, during a compliance review, or when adding a new dependency.Votes: 0GitHub stars: 4
- Llm Output GateCI gate that refuses to ship if the prompt-eval golden set regresses past threshold or prompt-injection-test fails at HIGH severity. Use when wiring CI protection for prompt changes, so a worse or unsafe prompt can't merge.Votes: 0GitHub stars: 4
- Local LlmHand off bulk transformation work (summarize, classify, extract, first-pass) to a local LLM via Ollama. Use when the task is high-volume and doesn't need Claude-grade reasoning, to save cost and latency.Votes: 0GitHub stars: 4
- Logging ScaffoldOne-shot scaffold dropping in a structured logger for the project's primary backend language, wiring JSON output, sensible defaults, and example call sites. Use when a backend lacks structured logging and needs it set up.Votes: 0GitHub stars: 4
- Mcp As AgentWrap an MCP server as a yakOS agent so it shares the same dispatch surface as LLM specialists. Use when an MCP server takes structured input and returns structured output like a specialist you want to dispatch to.Votes: 0GitHub stars: 4
- Mockup ReviewStructured review of a mockup — info hierarchy, interaction states, responsive behavior, copy slots, design-token usage. Use when evaluating a static design/mockup before implementation begins.Votes: 0GitHub stars: 4
- Monitor ScaffoldDrop in supervisor config + /healthz endpoint + restart runbook per service, per supervisor (systemd / pm2 / k8s / docker-compose). Use when services need health checks and process supervision set up.Votes: 0GitHub stars: 4
- Peer HandoffHand work to another developer cleanly in multi-dev co-pilot mode, packaging state, context, and next steps. Use when transferring an in-progress task to a peer so they can pick it up without losing context.Votes: 0GitHub stars: 4
- Peer SyncRead recent peer activity for a yakOS multi-dev session and summarize it for the lead. Use at session start in multi-dev mode, to catch up on what peers changed while you were away.Votes: 0GitHub stars: 4
- Perf Budget CheckCompare Core Web Vitals (LCP/INP/CLS) and bundle size against the project's perf budget, failing the PR if regressed. Use when reviewing a frontend change for performance impact or gating a release on perf.Votes: 0GitHub stars: 4
- Persona WriteScaffold a user persona document with required fields — demographics, goals, frustrations, behaviors, JTBD, devices, accessibility considerations. Use when grounding product or UX work in a concrete target user.Votes: 0GitHub stars: 4
- Phase CompleteVerify a phase's exit criteria are met before declaring it done. Use at the boundary of a project phase, before moving on or claiming completion, to catch unmet criteria.Votes: 0GitHub stars: 4
- Plan Quality EvalScore work/current/plan.md against a 6-dimension rubric using a 3-judge cross-vendor LLM panel and log a structured record. Use before specialists start work, to catch a vague or badly-decomposed plan while fixing it is cheap.Votes: 0GitHub stars: 4
- Postmortem WriteProduce a blameless postmortem — UTC timeline, impact, root cause, what worked / didn't, action items with owners and due dates. Use after an incident or outage is resolved, to capture root cause and follow-ups.Votes: 0GitHub stars: 4
- Pre CommitPre-commit verification — lint, secret scan, fast tests. Use right before committing, to catch lint errors, leaked secrets, and broken tests before they enter history.Votes: 0GitHub stars: 4
- Project InitBootstrap a new project's .claude/ config, complementing `yakos init`. Use when setting up a fresh project that needs its agent/skill/rule scaffolding in place.Votes: 0GitHub stars: 4
- Promote BranchPromote code between environments (dev → test → prod) via PR with a structured changelog, gated by the promotion-aware pre-push hook. Use when moving a tested change up the environment ladder toward production.Votes: 0GitHub stars: 4
- Prompt EvalRun a prompt against a golden dataset and report per-rubric regression vs the prior pinned version, not just an aggregate score. Use when changing a prompt, to confirm it didn't regress quality on known cases.Votes: 0GitHub stars: 4