erikuus
- 19 skills
- 0 followers
- 8 hours ago last updated
- ▌ Vau · erikuus bundleManage authenticated Rahvusarhiivi virtuaalne uurimissaal (VAU) content through the user's existing Chrome session, including editing Estonian or English articles, translating the latest Estonian news into English, and managing feedback. Invoke this skill explicitly with `$vau` when those workflows are needed.
- ▌ Autoreview · erikuus bundleStructured Codex, Claude, Amp, Pi, or Kimi code review when explicitly requested.
- ▌ Loop Creator · erikuus bundleCreate simple heartbeat candidate loops with exactly one input file, one criteria file, repeated candidate generation, repeated candidate evaluation, required creator/evaluator automations, and a static overview.html site. Use when the user wants a wit-loop or my-life-style loop where a creator produces candidates from input.md, an evaluator scores candidates against criteria.md, deterministic code archives results and updates the current best/shortlist, and an HTML overview shows progress. Do not use for complex executable transformation optimization or broad generic optimizer design.
- ▌ Review To Pr · erikuus bundleTake a finished Git topic branch through one isolated autoreview, verified in-scope repairs, focused tests, intentional commits, push, draft pull-request creation, and required-CI handoff without merging. Use when the user explicitly invokes review-to-pr after implementation is committed and wants the branch reviewed and published as a concise, human-reviewable PR instead of entering a multi-reviewer polling loop.
- ▌ Grill With Docs · erikuus bundleTurn a docs-heavy app planning or rewrite session into a reconstruction-grade, stack-neutral CONTEXT.md by grilling the user one question at a time. Use when the user has legacy architecture notes, UI descriptions, rationale, strengths and weaknesses, user feedback, migration intent, or product documents and wants to pressure-test the next version before implementation.
- ▌ Reprompt Design · erikuus bundleBuild project-root DESIGN.md and design-assets from screenshots, visual assets, styles, docs, and current app evidence as a stack-neutral visual source of truth for reimplementation. Use when the user wants to preserve or reconstruct an app's design, visual system, essential screens, responsive layouts, interaction states, or product-defining assets.
- ▌ Reprompt Refactor · erikuus bundleRebuild project-root SPEC.md from PLAN.md, DEVLOG.md, and current repo truth as a stack-neutral source-of-truth spec for reimplementation. Use when the user wants to synthesize the current implemented specification, rebuild SPEC.md, or reconcile the original plan with what the app actually became.
- ▌ Reprompt Harvester · erikuus bundleHarvest a finished coding thread into project-root DEVLOG.md for projects that keep PLAN.md, DEVLOG.md, and SPEC.md. Use when the user wants to finalize a thread into project docs, harvest a finished thread, or backfill a dev log from the current thread or archived Codex session .jsonl files.
- ▌ Refactor For Humans · erikuus bundleRefactor AI-generated or difficult repository code to reduce complexity and human cognitive load while preserving behavior. Use when Codex must inspect real code and simplify excessive files, tiny functions, thin wrappers, pass-through layers, shallow abstractions, caller-specific APIs, speculative patterns, duplicated rules, configuration flags, special cases, unclear ownership, tangled error handling, or broad AI rewrites; then compare designs, implement a focused refactor, and verify it with available tests.
- ▌ Sentry Fix Campaign · erikuus bundleTriage and remediate a set of related unresolved Sentry errors as one backward-compatible coding campaign on one Git branch with one eventual pull request. Use when Codex must inspect multiple errors for the current project, cluster duplicate or overlapping root causes, check whether earlier campaign commits already cover later issues, implement only missing fixes, add regression tests, and create intentional commits without separate implementation worktrees or branches per Sentry issue.
- ▌ Yii1 Sentry Logging · erikuus bundleInstall and verify Sentry logging in any Yii1 app using a frozen modern XSentryLogRoute baseline for PHP 7.2+ and a separate legacy PHP 5.6 branch, with Sentry REST API rule setup, Sentry plugin verification, and Gmail-based two-email proof workflow.
- ▌ Upgrade To Yii1 1 32 · erikuus bundlePlan and execute upgrades of legacy Yii 1.1.x apps to Yii 1.1.32 with a complete frontend/backend checklist and concrete code changes. Use when the user asks for "Upgrade to Yii1.1.32", "Yii 1.1.8 to 1.1.32", PHP 5.6 to PHP 8.x compatibility for Yii 1.1 apps, or jQuery/jQuery UI compatibility updates caused by Yii 1.1.32.
- ▌ Finalize Github Issue · erikuus bundleFinalize an implemented GitHub issue by reconstructing the completed work from the current analysis or development thread, verifying the user-visible solution at a supplied application URL in authenticated Chrome, capturing screenshots, and publishing a concise, complete issue comment through the connected GitHub app. Use only when explicitly invoked with an exact GitHub issue URL and application URL after implementation or review work is complete.
- ▌ Analyze Github Feature · erikuus bundleAnalyze a proposed feature before implementation by combining one GitHub issue, one supplied Markdown documentation file, the current project's code, optional documentation-directed Chrome observation of a test UI, and—when visual validation is useful—an isolated disposable Sites prototype; conduct an evidence-led Estonian clarification conversation in the issue, maintain a weekday working-hours same-task monitor, establish explicit participant consensus, and hand a technical plan back to the invoking Codex task without implementing the real feature. Use only when the user explicitly invokes `$analyze-github-feature` with a GitHub issue URL and documentation path. Never invoke implicitly for ordinary feature requests.
- ▌ Facebook Deletion Requests · erikuus bundleProcess Facebook user-data deletion CSV files, resolve app-scoped IDs through the VAU API, and create privacy investigation issues in a caller-specified GitHub repository. Use only when the user explicitly invokes $facebook-deletion-requests.
- ▌ Phoenix Multi Column Setup · erikuus bundleCreate and bootstrap Phoenix LiveView apps with a multi-column layout pattern. Automatically sets up the app, installs shared layout assets, bootstraps a local `mix multi_column.generate` pipeline, and generates auth-aware or no-auth navigation shells from a DSL spec.
- ▌ Toimeta Eestikeelset Teksti · erikuus bundleToimeta AI loodud või inglise keelest tõlgitud eestikeelset tarbeteksti nii, et see oleks loomulik, korrektne, sidus, tähendustäpne ja sihtrühmale sobiv. Kasuta, kui palutakse eemaldada tekstist AI-likkus või tõlkekeele jäljed, teha keele- või sisutoimetamist, kontrollida tõlget originaali alusel, ühtlustada stiili ja termineid, parandada veeb-, turundus-, äri- või teabeteksti või kohandada seda Eesti lugejale.
- ▌ Refactor Liveview For Humans · erikuus bundleRefactor AI-generated or difficult Elixir Phoenix LiveView code to reduce human cognitive load while preserving behavior. Use when Codex must inspect, diagnose, review, or simplify LiveViews, LiveComponents, function components, HEEx templates, contexts, changesets, routes, PubSub flows, streams, assigns, event handlers, hooks, helpers, tests, or generated component collections; especially for lifecycle duplication, unclear state ownership, misplaced domain or UI logic, excessive helpers, speculative abstractions, duplicated markup, unused artifacts, tangled params or navigation, and broad AI rewrites.
- ▌ Evaluate Maintainability For Humans · erikuus bundleEvaluate how maintainable an AI-generated repository is for a repository-new human by running live maintenance probes in isolated disposable worktrees, withholding location hints, interviewing the developer about navigation and cognitive load, inspecting the resulting change and change reasoning, and producing an evidence-backed complexity diagnosis and refactor handoff. Use when Codex must measure change friction before or after a refactor, diagnose unclear ownership, duplication, coupling, obscurity, change amplification, cognitive load, or unknown-unknown risk, or compare whether one design is easier for developers to modify. Supports any language or framework and live-change, read-only walkthrough, retrospective, and before/after evaluation modes.