Claude Skills by Roxabi

Walk through items 1-by-1 (findings, tasks, issues, TODOs). Triggers: "one by one" | "walk through" | "1b1" | "process each item" | "review one by one" | "go through each" | "handle them one at a time" | "step through".

Compress agent/skill definitions using math/logic notation. Triggers: "compress" | "compress skill" | "compress agent" | "compress context" | "shorten this" | "make it formal" | "use formal notation" | "expand notation" | "lint notation" | "derive pattern from skills".

Initialize the CV plugin — creates data directories and starter cv_data.json template. Triggers: "cv-init" | "init cv" | "setup cv".

Generate, adapt, and update CVs from structured data stored in cv_data.json. Triggers: "cv" | "generate cv" | "generate my cv" | "adapt cv" | "update cv" | "tailor cv" | "make my resume" | "update my resume" | "write my cv" | "adapt my resume for" | "tailor my cv for".

**Archived:** 2026-08-22. Skill removed from the active plugin; use `/frame` for problem framing or `/analyze` for architecture exploration instead. --- --- name: clarify argument-hint: '["topic" | --issue <N> | --resume]' description: Intent-first architecture recap — explain what we are really solving (intent → biz-arch → UX flows → data flow per layer → use cases × layers → open intent Qs); defer technical implementation until approved. Phase-agnostic, ephemeral, no artifact. Triggers "cla...

Create/list Architecture Decision Records. `--axial` elicits the Axis of Decomposition ADR.

Red-team / devil's advocate on analyses, proposals, architecture, ideas, specs, or plans — attack assumptions, kill vacuous claims, surface bypass and failure modes. Triggers: "adversarial" | "red team" | "devil's advocate" | "attack this design" | "kill this idea" | "stress test this" | "what breaks this" | "adversarial review".

Constructive expert advisory on analyses, proposals, architecture, ideas, specs, or plans — strengthen, prioritize, surface blind spots as recommendations (not red-team). Triggers: "advisory" | "second opinion" | "advise on this" | "strengthen this" | "expert advice" | "advisory review" | "what would you improve".

Deep technical analysis — explore existing code, risks, shapes. Triggers: "analyze" | "technical analysis" | "how deep is it" | "deep dive" | "investigate this" | "analyze this feature" | "what are the risks" | "explore the codebase" | "look into this" | "explain the architecture" | "what is the architecture" | "explain from intent down".

Set up CI/CD — GitHub Actions workflows, TruffleHog, Dependabot, pre-commit hooks, marketplace plugins. Triggers: "ci setup" | "setup ci" | "configure ci" | "setup hooks" | "setup github actions".

Watch a CI run with live status + emoji dashboard; dump logs on failure; watch auto-merge if CI green + auto-merge enabled + "reviewed" label. Triggers: "watch ci" | "ci watch" | "watch the ci" | "watch run" | "monitor ci".

Audit and clean CLAUDE.md, memory, skills, and rules — resolve every finding (fix/promote/relocate/delete), track recurrences, target bloat=0. Triggers: "cleanup context" | "context audit" | "clean memory" | "drain memory" | "prune memory" | "audit memory" | "consolidate rules" | "spa day" | "memory audit".

Clean git branches/worktrees/remotes after merge-status verification; sweep stuck pipeline labels and orphan CI runs. Triggers: "cleanup" | "clean branches" | "cleanup worktrees" | "remove stale branches".

Health check — dev-core config, GitHub, labels, workflows, branch protection, CI. Not host /checkup.

Execute plan — setup worktree, spawn agents, write code + tests. Triggers: "dev-implement" | "implement" | "build this" | "execute plan" | "start coding" | "write the code" | "code this up" | "let's build it" | "build it out" | "/R-dev-implement". Not the host native /implement.

Initialize project — env-setup, ci-setup, release-setup + axial ADR. Not host /init.

Implementation plan — tasks, agents, file groups, dependencies. Triggers: "dev-plan" | "plan this" | "implementation plan" | "break it down" | "plan this feature" | "how should we build this" | "make a plan" | "create a plan" | "break this down into tasks" | "task breakdown" | "/R-dev-plan". Not the host native /plan.

Multi-domain code review (agents + Conventional Comments → findings + verdict). Triggers: "dev-review" | "code review" | "review changes" | "review PR #42" | "check my code" | "review my changes" | "review this PR" | "do a code review" | "review the diff" | "look at my code" | "/R-dev-review". Not the host natives /review or /doctor.

Workflow orchestrator — single entry point for the full dev lifecycle. Triggers: "dev" | "start working on" | "work on issue" | "work on #" | "develop" | "pick up issue" | "tackle issue" | "let's work on".

Sync project docs after a code change — stale references, affected sections.

Set up local dev environment — stack.yml, AGENTS.md Critical Rules, docs scaffolding, LSP. Triggered by /R-dev-init or standalone /R-env-setup. Triggers: "env setup" | "setup environment" | "scaffold rules".

Apply review findings — auto-apply high-confidence, 1b1 for rest, then batch-apply. Triggers: "fix findings" | "fix review" | "apply fixes" | "fix these" | "apply review comments" | "apply the review" | "fix the review issues" | "address review feedback" | "fix PR comments".

Problem framing — capture problem, constraints, scope, tier. Triggers: "frame" | "frame this" | "what's the problem" | "define the problem" | "scope this out" | "define the scope" | "what are we solving" | "problem statement" | "restate what we are solving" | "recap the issue" | "intent first" | "step back and explain" | "explain it properly" | "restructure the answer".

Structured interview → brainstorm | analysis | spec (with promotion). Triggers: "interview" | "brainstorm" | "let's brainstorm" | "think through this" | "help me brainstorm" | "let's think this through" | "explore ideas".

Create/update PRs with Conventional Commits title, issue linking & guard rails. Triggers: "create PR" | "open PR" | "submit PR" | "update PR" | "/R-pr --draft" | "open a pull request" | "make a PR" | "open pull request" | "submit a pull request" | "create a draft PR" | "raise a PR".

Promote staging→main — pre-flight, version bump, changelog, PR and tag.

Audit README, CONTRIBUTING, and plugin READMEs against the developer-tool pattern.

Drift-check an issue before work begins — fails fast when code has evolved (git-drift, symbol-missing, dep-resolved). Triggers: "recheck" | "is this issue still valid" | "check drift" | "check issue staleness".

Set up commit standards and release automation — Commitizen, commitlint, semantic-release, Release Please, Lefthook/Husky. Triggers: "release setup" | "setup releases" | "commit standards" | "setup release automation".

Bootstrap OSS community health files (CONTRIBUTING, LICENSE, SECURITY, templates).

Populate scaffolded architecture/standards docs from AGENTS.md and the codebase.

Create + link feature branch to issue, then check out worktree. Triggers: "setup worktree" | "create worktree" | "prepare workspace" | "bootstrap branch".

Meta-orchestrator to land ready code — commit → PR → code-review → fix loop → reviewed label + ci-watch. Triggers: "ship" | "ship this" | "land this" | "land the PR" | "ship PR" | "submit for merge" | "get this merged" | "ship my branch".

Solution spec — acceptance criteria, breadboard, slices. Triggers: "write spec" | "spec this" | "solution design" | "what will we build" | "design the solution" | "acceptance criteria" | "define acceptance criteria" | "spec it out" | "write the spec".

Fill `.dev/stack.yml` via guided discovery.

Generate or run unit, integration, and Playwright e2e tests.

Run all quality gates (lint, typecheck, test, env, i18n, license; optional falsify on --full) and produce a structured pass/fail report. Triggers: "validate" | "quality check" | "pre-push check" | "are we green".

Extract and store invoice details from documents — vendor, date, amount, line items, tax, payment terms. Saves structured JSON to vault. Triggers: "get-invoice-details" | "extract invoice" | "parse invoice" | "invoice details" | "read invoice" | "analyze this invoice" | "process this invoice" | "analyse cette facture" | "extract this invoice" | "get invoice data".

Index a repository into a GitNexus knowledge graph. Triggers: "index repo" | "analyze codebase" | "gitnexus analyze" | "build knowledge graph" | "create graph index".

360-degree view of a code symbol — callers, callees, processes it participates in. Triggers: "symbol context" | "show callers" | "gitnexus context" | "who uses this" | "what does X call".

Blast radius analysis — what breaks if you change a symbol. Triggers: "impact analysis" | "what breaks if" | "blast radius" | "gitnexus impact" | "is it safe to change" | "dependencies of".

List all GitNexus-indexed repositories. Triggers: "list repos" | "gitnexus list" | "what repos are indexed" | "show indexed repos".

Search the knowledge graph for execution flows related to a concept. Triggers: "search graph" | "find execution flow" | "gitnexus query" | "how does X work" | "find code related to".

Check GitNexus index status for current repo — freshness, symbol count, staleness. Triggers: "check index" | "gitnexus status" | "is graph stale" | "index status".

Generate AI image prompts with visual identity and style consistency — reads brand charter, applies artistic styles, outputs multiple prompt variants. Triggers: "image-prompt" | "generate image prompt" | "image prompt" | "prompt for image" | "visual prompt" | "generate a prompt for Grok" | "generate a prompt for Midjourney" | "generate a prompt for flux" | "image generation prompt" | "create image prompt" | "prompt for dall-e".

Triage/create GitHub issues — set size/priority/lane/type labels, manage dependencies & parent/child. Run before `/feature` picks a ticket up. Triggers: "triage" | "create issue" | "set size" | "set priority" | "blocked by" | "set parent" | "child of" | "sub-issue" | "file an issue" | "log a bug" | "open an issue" | "file a bug" | "add issue" | "new issue" | "set lane" | "set type".

Initialize the LinkedIn Apply plugin — create vault candidate profile, install Playwright, and set up LinkedIn browser session. Triggers: "linkedin-apply-init" | "init linkedin apply" | "setup linkedin apply" | "linkedin apply setup".

Scrape a LinkedIn job offer and score it against your profile using Claude LLM — produces APPLY / REVIEW / SKIP decision with 4-dimension scores. Triggers: "linkedin apply" | "analyze job" | "job match" | "linkedin-apply" | "score this job" | "should i apply for this job" | "evaluate this job posting" | "is this job a good fit" | "check this job offer" | "rate this job".

Generate engaging LinkedIn posts with best practices, visual identity, and vault integration. Triggers: "linkedin post" | "write linkedin" | "generate linkedin" | "linkedin content" | "post for linkedin" | "create a linkedin post" | "draft a linkedin post" | "write a post for linkedin" | "linkedin update".

OMP-only — grill then validate spec, then plan→impl→PR→review→reviewed→watch merge. Not host /dev.