intrusive-memory

Ship one or more iOS/macOS App Store apps from a single repo through development → main → TestFlight, then promote to the App Store by tagging a release candidate (v<version>-rc.<k>) on main and merging that tag into a release branch. Handles repos that ship several apps at once (e.g. an iOS app + a macOS app, each its own App Store Connect record/scheme). Orchestrates the development→main PR cycle (which auto-ships to TestFlight via Xcode Cloud), then the RC-tag → release promotion that uploads to App Store Connect, then submission. Xcode Cloud only builds/archives/uploads; GitHub CI owns all test gating. Use when asked to ship, release, cut a release candidate, promote to the App Store, run release QA, or push iOS/macOS apps to TestFlight or the App Store.

Ship and release Swift library versions by bumping the version on development, merging the PR once CI passes, then tagging and creating a GitHub release from main. Accepts an optional version-bump argument (`patch`/`minor`/`major`) or explicit semver (`1.2.3`) to run unattended.

Audit a Swift test suite across five dimensions and produce a single markdown findings report with concrete recommendations. Run multiple passes that each look at the test suite from a different angle — repetition, superfluity, coverage gaps (using real xccov data), flaky-in-CI risk, and CI gating of performance tests. Use this skill whenever the user asks to audit, analyze, review, critique, or clean up their tests, including phrases like "test analysis", "audit my tests", "review the test suite", "are there redundant tests", "find flaky tests", "test coverage gaps", "which tests can I delete", "are my tests good", or whenever they want a holistic read on Swift test-suite health. Triggers on Swift Package Manager / Xcode projects with XCTest or swift-testing test targets. Defer to Makefile targets when present; never run `swift test` or `swift build`.

Generate or refresh a graphify knowledge-graph codemap of the current repo, gitignore the machine-specific/regenerable pieces, and commit only the portable map artifacts. Use whenever the user wants to build, update, refresh, regenerate, sync, or commit the codemap / graphify map / code graph for a project — phrases like 'update the codemap', 'refresh the graphify map', 'rebuild the code graph', 'commit the codemap', 'regenerate graphify-out', 'sync the codemap before I push', or any variation of keeping a checked-in graphify-out/ map current. Handles first-time generation and incremental updates, enforces the correct .gitignore split (absolute-path markers + cache + manifest.json stay out; graph.json/html/report stay in), untracks anything now-ignored, ensures AGENTS.md references the codemap, and makes one clean commit. Optionally pushes.

Finalize the pull request for the current branch — find the existing (usually draft) PR, regenerate its title and body from the branch's commits, fix the base branch if it's wrong, and mark it ready for review. Falls back to creating a new PR only when none exists. Use this skill whenever the user asks to "create a PR", "open a PR", "make a pull request", "PR this", "send it up", "mark it ready", "ready for review", "finalize the PR", or any variation that implies promoting the current branch into review — including casual phrasings like "now PR it" or "go ahead and ship the PR". The skill picks the base branch using the project's tiered convention (mission/* → development, development → main, everything else → development), enforces standard safety checks (clean working tree, not on main, push first), and writes a Summary + Test plan body from the commits on the branch. Prefer this skill over running `gh pr create` / `gh pr ready` directly so the routing and safety rules apply consistently.

Plan and execute sorties with sergeant precision. Give each agent ONE clear, measurable goal. Pre-execution commands (breakdown, refine + 4 subcommands) create and refine an EXECUTION_PLAN.md from requirements. Refine performs 4 passes: atomicity/testability, prioritization, parallelism (up to 4 sub-agents, builds only by supervisor), and open questions. Execution commands (start, resume, status, stop, killall) orchestrate sortie agents with lean context and crystal-clear objectives. THE RITUAL (name-feature) generates humorous military operation names. Post-mission flow runs automatically after the last sortie completes: (test-cleanup) prunes tests added during the mission that cannot run reliably in CI, then (brief) harvests lessons and renders an explicit ROLLBACK | KEEP | PARTIAL_SALVAGE verdict, then auto-triggers (clean) → /organize-agent-docs to archive every mission artifact in the project root into docs/<complete|incomplete>/<mission name>/.

Gold-standard markdown organization for any repository. Sorts every markdown file into one of three buckets — FOUNDATIONAL (AGENTS.md/CLAUDE.md/GEMINI.md/README.md/CHANGELOG.md/LICENSE at root), MISSION (mission-supervisor artifacts; lifecycle tracked via frontmatter `state:` and routed to root, docs/complete/, or docs/incomplete/), or EXTRANEOUS (everything else → docs/). Maintains links in FOUNDATIONAL files after moves and stamps an `updated:` date when content changes. Use when the user asks to "organize agent docs", "clean up markdown", "tidy the repo", "archive mission files", "separate agent instructions", "restructure AGENTS.md", "sort docs", or whenever a project's markdown files are mixed between root and docs/. This skill is the single owner of repo-level markdown moves; mission-supervisor delegates its `clean` step here.

Audits and automatically fixes Swift Package Manager library packages. Removes Package.resolved from git tracking, ensures the sibling dependency pattern is in place via /toggle-sibling-libraries, and bumps every intrusive-memory/* dep to its latest published release. Explicitly invoked only — does not auto-trigger.

Toggle a Swift Package.swift between the local-sibling-checkout development pattern and a clean remote-only release pattern. Use this skill whenever you need to flip Package.swift in or out of the `sibling(...)` helper-function pattern — for example, before tagging a release the dependencies must be pinned to remote-only, and on a fresh `-dev` development cycle they must go back to sibling-aware. Also trigger on phrases like "switch to remote deps", "remove the sibling helper", "restore sibling pattern", "make Package.swift release-ready", "flip to dev-mode deps", or any time the user references the `sibling()` helper or the `useLocalSiblings` constant. Only intrusive-memory/* dependencies participate in the sibling pattern; all other dependencies are left untouched.

Plan and execute sprints with sergeant precision. Give each agent ONE clear, measurable goal. Pre-execution commands (breakdown, analyze, prioritize, evaluate) create and refine an EXECUTION_PLAN.md from requirements. Execution commands (start, resume, status, stop, killall) orchestrate sprint agents with lean context and crystal-clear objectives. THE RITUAL (name-feature) generates humorous military operation names.

Use macOS text-to-speech via the `say` command for voice feedback, audio narration, and spoken output.

UI/UX design intelligence. 50 styles, 21 palettes, 50 font pairings, 20 charts, 8 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app, .html, .tsx, .vue, .svelte. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient.

Generate audio and transcripts for podcast episodes from Fountain screenplays using Produciesta CLI. When asked to "generate podcast audio", "create audio plan", "cast voices", "prepare transcripts", or working with podcast projects that have episodes/*.fountain files, this skill analyzes characters, assigns voices, writes PROJECT.md cast list, generates audio, and prepares transcripts for web deployment.

Organize project agentic documentation into universal (AGENTS.md) and agent-specific files (CLAUDE.md, GEMINI.md, etc.). Use when asked to "organize agent docs", "separate agent instructions", "restructure AGENTS.md", or when a project has agent documentation that mixes universal and tool-specific content.

Ship and release Swift library versions by bumping the version on development, merging the PR once CI passes, then tagging and creating a GitHub release from main

Release a merged PR by tagging and creating a GitHub release. Use when asked to "release", "tag", or "create a release".

Wire a repository's model-dependent integration test(s) to actually run in CI against cached SwiftAcervo models, the standardized way. Use this skill whenever the user wants to fix, enable, add, or standardize integration tests that need a downloaded ML model in GitHub Actions — phrases like "make the integration tests run in CI", "my model tests are disabled in CI", "cache the models in CI", "add the integration test to the pipeline", "standardize how we test real inference", or "the generation/TTS/inference test never runs on CI". It vendors a credential-free CDN primer + cache-key script into .github/scripts/, adds/updates an integration-tests workflow that restores models via actions/cache and primes on miss, wires the TEST_RUNNER_ACERVO_* env so xcodebuild forwards the cache into the xctest runner, and normalizes the test's gate to model-presence (XCTSkipUnless / @Test(.enabled(if:))). Input is the specific integration test (file/target/suite) when it can't be derived from repo context. Works one repo at

Run a queue of `acervo ship` model uploads sequentially — launch one, wait for it to complete and verify, then launch the next, until the list is exhausted. Use this skill whenever the user wants to ship, mirror, publish, or upload multiple models to the intrusive-memory CDN in one go — phrases like "ship these models", "batch upload", "queue up these acervo ships", "publish this list of models", "run acervo on all of these", "process this models.txt", or any time the user hands over more than one HuggingFace `org/repo` id and wants them shipped one after the other. Also triggers on follow-up checks like "how's the queue going", "what's left in the queue", or "is the batch done". Builds on the single-model `acervo-download-ship` skill and reuses its launcher; the queue layer just handles sequencing, per-item finalization, and skip-on-failure progression. Integrates with `/loop` dynamic mode so the agent isn't pinned waiting on hour-long uploads.

Launch a long-running `acervo ship` model download in a detached shell so the agent isn't blocked while multi-gigabyte models stream from HuggingFace to the intrusive-memory CDN. Trigger this skill whenever the user asks to ship, mirror, publish, upload, push, or download a model with acervo — phrases like "acervo ship X", "ship mlx-community/Y", "publish this model to the CDN", "mirror Z to R2", or "kick off the download for org/repo". Also use it to check on or report status for any acervo download already in flight ("how's the download", "is it done yet", "check the acervo log"). The skill captures verbose output to a log file, enforces single-download-at-a-time, and integrates with `/loop` dynamic mode so checks happen automatically without the agent having to wait.