llamenos-platform

Use when adding encryption for a new data type, implementing envelope encryption, working with HPKE key wrapping, managing crypto labels, or handling multi-recipient decryption. Also use when the user mentions "encrypt", "E2EE", "envelope", "HPKE", "forward secrecy", "crypto label", "admin envelope", "content key", or needs to understand how data is encrypted at rest.

Guide JSON Schema and protocol changes through the multi-platform codegen pipeline in the Llamenos monorepo. Use this skill when modifying files in packages/protocol/schemas/, changing crypto-labels.json, updating wire formats, adding new message types, or when the user mentions "schema change", "protocol update", "codegen", "wire format", "new message type", "crypto label", "domain separation", "add a field", "new field", "envelope format", or "shared types". Also use when the user describes adding or modifying data structures that need to work across all platforms (desktop, iOS, Android, worker) — this implies a protocol change. If a change touches encrypted payload formats or cross-platform type definitions, this skill applies. Schema changes have blast radius across every platform — this skill ensures nothing is missed.

Process security audit findings into dependency-ordered implementation epics for the Llamenos monorepo. Use this skill when conducting a security audit, reviewing security findings, processing vulnerability reports, creating security fix epics, or when the user mentions "security audit", "vulnerability", "findings", "security review", "pen test results", "audit round", "security issue", "security bug", "exploit", "insecure", "hardening", or "timing-unsafe". Also use when the user describes specific security problems inline — like missing auth checks, unsafe crypto operations, or data exposure — even without using the word "audit" or "vulnerability". If the user pastes or describes 2+ security issues that need triage and remediation, this skill applies. Covers all platforms: Worker backend, desktop (Tauri), iOS (SwiftUI), Android (Kotlin/Compose), crypto (Rust), CI/CD, and infrastructure.

Guide BDD test writing for the Llamenos monorepo. Use this skill when writing Gherkin feature files, creating backend or desktop step definitions, debugging test failures, or when the user mentions "BDD", "feature file", "Gherkin", "step definition", "backend BDD", "test:backend:bdd", "behavioral test", "write tests first", "add test coverage", "tests broke", "fix tests", or "E2E testing". This skill supplements superpowers brainstorming and plan execution — it does not replace them. Use it for BDD-specific implementation detail, not as a workflow entry point.

Guide porting features across Desktop (Tauri/React), iOS (SwiftUI), and Android (Kotlin/Compose) in the Llamenos monorepo. Use this skill when implementing a feature that exists on one platform and needs to work on another, when the user mentions "port to iOS", "Android version", "feature parity", "cross-platform", "implement on all platforms", "build the iOS/Android/mobile version", "same thing on Swift/Kotlin/native side", or describes a feature that "doesn't exist on [platform] yet". Also use when the user says "the desktop version works, now build it for mobile", asks to "replicate" or "mirror" a feature across platforms, or when creating epics for mobile feature parity. If a user describes building a screen or feature and mentions a specific target platform while referencing an existing implementation on another platform, this skill applies. It ensures consistent architecture, shared types, proper crypto integration, and correct i18n string handling on each platform.

iOS build, test, and debugging workflows for the Llamenos project. Covers SSH to Mac M4, xcodebuild commands, simulator selection, XCFramework rebuilds, and UniFFI binding sync. Use when building iOS, running XCUITests, debugging crypto FFI issues, or troubleshooting Xcode/SPM problems.

Use when deploying the application, bumping versions, cutting releases, configuring CI/CD, managing Docker/Helm/Ansible deployments, or troubleshooting deployment issues. Also use when the user mentions "deploy", "release", "version bump", "Docker Compose", "Helm", "CI/CD", "wrangler", or needs to understand the deployment topology and release process.

Use when adding new Tauri IPC commands, modifying desktop crypto operations, updating the Rust CryptoState, or fixing Playwright test failures related to missing IPC mocks. Also use when the user mentions "IPC command", "platform.ts", "CryptoState", "tauri invoke", "generate_handler", or needs to understand the 4-layer IPC chain from Rust to test mocks.

Use when adding a new telephony or messaging provider, implementing webhook validation, working with IVR flows, or modifying call routing. Also use when the user mentions "telephony adapter", "messaging adapter", "webhook validation", "TwiML", "NCCO", "Plivo XML", "Asterisk", "SMS", "WhatsApp", "Signal", "provider", or needs to understand the multi-provider abstraction layer.

Orchestrate test runners, simulators, emulators, and backend services across all platforms in the Llamenos monorepo. Use this skill when setting up test environments, launching simulators/emulators, starting Docker Compose backends, running cross-platform test suites, debugging CI failures, or when the user mentions "run tests", "test:all", "test:changed", "simulator", "emulator", "Docker Compose tests", "CI failure", "test environment", "parallel tests", "verify everything", "codegen guard", "build failing", or "build error". Also use when the user needs to run verification after making changes across multiple platforms, when configuring new test environments, debugging why tests pass locally but fail in CI, or when test infrastructure itself is broken (stuck simulators, port conflicts, stale codegen). Distinct from bdd-feature-development which diagnoses test logic failures — this skill handles the infrastructure and orchestration layer. Also covers backend BDD test orchestration against Docker Compose.

Use when cutting a release, signing desktop artifacts, promoting builds to production, uploading to RustFS, generating Tauri updater manifests, or when the user says "/release", "sign release", "promote build", "cut a release", or "publish desktop". Walks through the complete offline signing workflow with strict verification gates.

Guide dependency and tooling upgrades across all platforms in the Llamenos monorepo. Use this skill when upgrading Gradle/AGP/KSP (Android), Xcode/SPM (iOS), Cargo/Rust (crypto), Bun/npm packages (desktop/worker), or any cross-cutting dependency. Also use when the user mentions "upgrade", "update dependencies", "bump version", "new version of", "migrate to", "breaking change from upstream", "outdated packages", "bun audit", "cargo outdated", "security advisory", "npm vulnerability", or when Dependabot PRs need processing. Use when the user asks to check for or fix dependency vulnerabilities, wants to know if packages are up to date, or when a new major version of a framework (Tauri, Compose, SwiftUI) is released and the user wants to evaluate or apply the upgrade. This skill prevents the most common upgrade failures: breaking API changes, missed platform-specific migration steps, and untested regressions.

Manage internationalization strings across 13+ locales and 3 platforms (desktop, iOS, Android) in the Llamenos monorepo. Use this skill whenever adding, renaming, moving, or removing i18n strings, when the user mentions "i18n", "localization", "translation", "locale", "strings", "add a string", "new language", "all languages", "translate", "multilingual", or when modifying any file in packages/i18n/. Also use when the user says something needs to work "in all languages", asks to "add support for [language name]" (e.g., "add Swahili", "add Japanese"), or mentions string keys like t('key') or R.string.key. Use when running codegen or validation for i18n, fixing missing string references, or when any feature work implicitly requires new user-facing text. This skill prevents the most common i18n bugs: missing locale propagation, wrong key casing, stale codegen output, and broken platform references.