tacogips

Use when implementation work in a rielflow project changes behavior, adds functionality, or fixes bugs and the user has not explicitly asked to avoid workflows. Routes the work through the packaged `claude-code-design-and-implement-review-loop` workflow, including design/plan alignment, implementation, review, user-facing documentation refresh, commit-message generation, and built-in git commit/push steps.

Use when implementation work in a rielflow project changes behavior, adds functionality, or fixes bugs and the user has not explicitly asked to avoid workflows. Routes the work through the packaged `codex-design-and-implement-review-loop` workflow, including design/plan alignment, implementation, review, user-facing documentation refresh, commit-message generation, and built-in git commit/push steps.

Use when requested work should first be converted into a project-scope Rielflow workflow and then executed through Rielflow. Applies to workflow-first implementation, review, refactoring, documentation, or maintenance tasks where no existing package workflow exactly fits, and to requests that explicitly say to work through a project-scope workflow.

Use when requested work should first be converted into a project-scope Rielflow workflow and then executed through Rielflow. Applies to workflow-first implementation, review, refactoring, documentation, or maintenance tasks where no existing package workflow exactly fits, and to requests that explicitly say to work through a project-scope workflow.

Use when running, editing, or troubleshooting the packaged claude-code-deepdesign rielflow workflow for design-doc specification authoring with deep and broad review loops.

Use when installing a rielflow workflow bundle from a public GitHub directory URL with workflow checkout. Covers project/user scope destinations, duplicate handling, overwrite behavior, registry metadata, validation, and checkout troubleshooting.

Use when running a generic Claude Code Opus 4.8 workflow that creates and reviews a goal, creates and reviews an in-session plan without plan files, performs the work, and reviews whether the goal is achieved before looping back to work or planning.

Use when planning, running, monitoring, or troubleshooting the packaged rielflow divide-and-conquer refactoring workflow. Applies to `claude-code-refactoring-divide-and-conquer`, duplicate-scavenge refactoring, concurrent slice review via `claude-code-refactoring-slice-review`, generated refactoring implementation plans, implementation/review loop handling, blocked verification, dependency-owned agent mocks, and final plan archival.

Use when requested work should first be converted into a project-scope Riela workflow and then executed through Riela. Applies to workflow-first implementation, review, refactoring, documentation, or maintenance tasks where no existing package workflow exactly fits, and to requests that explicitly say to work through a project-scope workflow.

Use when creating, modifying, reviewing, or validating riela workflow bundles. Applies to step-addressed workflow.json authoring, node payload files, prompt files, built-in node add-ons, cross-workflow transitions, manager/worker routing, and portable workflows under a workflow root.

Use when creating, validating, or running temporary riela workflows from inline JSON or JSON files without installing a project/user-scope workflow bundle. Applies to --workflow-json, --workflow-json-file, embedded prompt-only payloads, temp workflow payload logs, and dry-run smoke checks.

Use when creating, modifying, reviewing, or validating riela workflow bundles. Applies to step-addressed workflow.json authoring, node payload files, prompt files, built-in node add-ons, cross-workflow transitions, manager/worker routing, and portable workflows under a workflow root.

Use when maintaining the riela-packages registry for package updates, release preparation, publishing registry changes, refreshing riela-package.json checksum and integrity digests, validating packaged workflows, updating README package listings, or doing pre-push release checks for packages under packages/*.

Use when maintaining a riela-packages registry for package updates, release preparation, publishing registry changes, refreshing riela-package.json checksum and integrity digests, validating packaged workflows, updating README package listings, or doing pre-push release checks for packages under packages/*.

Use when maintaining a riela-packages registry for package updates, release preparation, publishing registry changes, refreshing riela-package.json checksum and integrity digests, validating packaged workflows, updating README package listings, or doing pre-push release checks for packages under packages/*.

Use when configuring, running, inspecting, or troubleshooting riela auto-improve supervision. Applies to workflow run --auto-improve, nested supervisor/superviser workflows, supervision policies, stall detection, remediation budgets, workflow patch modes, targeted reruns, supervision state, nestedSuperviserDriver, and GraphQL/library parity for supervised execution.

Use when configuring, running, inspecting, or troubleshooting divedra auto-improve supervision. Applies to workflow run --auto-improve, nested supervisor/superviser workflows, supervision policies, stall detection, remediation budgets, workflow patch modes, targeted reruns, supervision state, nestedSuperviserDriver, and GraphQL/library parity for supervised execution.

Use when configuring, validating, emitting, serving, listing, replaying, or troubleshooting divedra external event sources. Applies to .divedra-events, events validate/emit/serve/list/replay, webhook/chat/cron bindings, event-to-workflow input mapping, supervised event execution, mock scenarios for events, and event receipt inspection.

Use when implementation work in this repository changes behavior, adds functionality, or fixes bugs and the user has not explicitly asked to avoid workflows. Routes the work through the project-local divedra workflow `.divedra/workflows/design-and-implement-review-loop`, including design/plan alignment, implementation, review, user-facing documentation refresh, commit-message generation, and built-in git commit/push steps.

Use when implementing or issuing divedra manager control-plane actions. Applies to divedra graphql, GraphQL manager mutations, sendManagerMessage, retry-step, replay-communication, optional-step execute/skip decisions, manager session auth, idempotency keys, scoped attachments, and avoiding freeform-only privileged control.

Use when using, reviewing, or implementing divedra node add-ons. Applies to workflow.nodes[].addon, built-in divedra add-ons, local scoped add-ons, third-party add-on resolvers, addon.config, addon.inputs, addon.env, node add-on validation, and add-on-backed worker nodes.

Use when diagnosing failed, paused, stalled, timed out, or surprising divedra workflow executions. Applies to session status/progress/logs/export, runtime artifacts, node execution records, workflow validation failures, backend invocation errors, timeout failures, GraphQL inspection, event receipts, and artifact/session-store path issues.

Use when explaining or operating the divedra terminal UI as an end user. Applies to launching the TUI/server UI, workflow overview/status browsing, pane navigation, selected workflow/session inspection, keyboard movement expectations, and operator workflows. Do not use for implementing TUI code changes; use tui-navigation-guardrails for code modifications.

Use when organizing, deduplicating, or refactoring divedra workflow bundles. Applies to reducing duplicate workflows, extracting reusable called workflows through cross-workflow transitions, using called-workflow composition, consolidating repeated prompts/nodes/add-ons, reviewing historical runtime artifacts, and improving workflow structure or input/output contracts from observed execution failures.

Run mixed Tauri Rust and TypeScript lint, typecheck, formatting, and verification commands. Use when Codex is asked to lint, check, test, verify, or prepare Rust/TypeScript/Tauri changes in a Tauri template project, especially after editing src-tauri, src, package.json, Taskfile.yml, flake.nix, or agent lint scripts.

Use when building, validating, publishing, or tap-rendering Homebrew formula releases for this Go project, including scripts/build-homebrew-release.sh, scripts/render-homebrew-formula.sh, Taskfile Homebrew tasks, GitHub Release assets, and tap formula verification.

Use when setting up or verifying Apple Developer ID signing credentials, app-specific passwords, kinko secret storage, local keychain identities, notarytool credentials, or macOS notarization readiness for this Swift Homebrew Cask workflow without exposing credential values.

Use when building, validating, publishing, or tap-rendering Homebrew formula tarball releases for this Swift project, including scripts/build-homebrew-release.sh, scripts/render-homebrew-formula.sh, and task build:homebrew or homebrew:formula commands.

Use when building, signing, notarizing, validating, publishing, or tap-rendering macOS Homebrew Cask DMG releases for this Swift project, including Apple Developer ID signing, scripts/build-homebrew-cask-release.sh, scripts/render-homebrew-cask.sh, and release:homebrew-cask-local.

Use when executing tasks from implementation plans. Provides task selection, parallel execution, progress tracking, and review cycle guidelines.

Use when creating test plans from implementation and design documents. Provides test plan structure, test case tracking, and coverage guidelines.

Use when creating execution plans from design documents. Provides plan structure, status tracking, dependencies, and progress logging guidelines for investigation or implementation work.

Run mixed Tauri Rust and TypeScript lint, typecheck, formatting, and verification commands. Use when Codex is asked to lint, check, test, verify, or prepare Rust/TypeScript/Tauri changes in this repository, especially after editing src-tauri, src, package.json, Taskfile.yml, flake.nix, or agent lint scripts.

Use when setting up or verifying Apple Developer ID signing credentials, Apple app-specific passwords, kinko secret storage, or local macOS notarization readiness for chilla without recording credential values.

Use when Codex should validate chilla macOS desktop behavior, Tauri app launches, visible UI flows, or browser/app interactions after implementation or bug fixes, and Computer Use tools may be available.

Use when code changes in this repository should be validated by actually launching the local chilla app. After modifying runtime-affecting code, rebuild the debug app when needed and launch `target/debug/chilla` yourself instead of telling the user to run it.

Use when building, signing, notarizing, validating, or publishing macOS `.app` and `.dmg` release artifacts for this repository. Covers the Tauri macOS release config, local `task bundle-macos-dmg`, Apple signing/notarization inputs, GitHub Actions release workflow, and post-release tap follow-up.

Use when building and publishing the Nix tarball release artifacts for this repository with `nix build`. Covers flake target selection, tarball packaging, installer compatibility, GitHub release creation, and Homebrew tap updates for the current tarball-based cask path.

Use when writing, reviewing, or refactoring Rust code. Provides type safety patterns, error handling, project layout, and async programming guidelines.

Use when writing, reviewing, or refactoring TypeScript code. Provides type safety patterns, error handling, project layout, and async programming guidelines.

Execute QraftBox release operations end-to-end using Taskfile tasks, including GitHub Release and npm publish. Use when users ask to release, publish a version, or run post-merge release operations.

Use when writing TypeScript code that interacts with dependencies, handles credentials, executes child processes, or manages configuration. Provides Shai-Hulud supply chain attack countermeasures at the code level including safe dependency usage, credential handling, subprocess hardening, and runtime integrity patterns.

Use when installing, updating, or auditing npm dependencies with Bun. Provides Shai-Hulud supply chain attack countermeasures including bunfig.toml hardening, lockfile verification, trustedDependencies management, and CI/CD pipeline security.

Use when creating, publishing, or maintaining npm packages with Bun. Provides Shai-Hulud supply chain attack countermeasures including npm token management, 2FA enforcement, provenance signing, trusted publishing via GitHub Actions, and pre-publish security checklists.

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

Use when creating or organizing design documents. Provides directory structure, file naming, and content guidelines for design specs and references.

Test-Driven Development with visual UI confirmation. Uses agent-browser for ad-hoc spot verification and Playwright for repeatable regression tests. Covers the full Plan-Red-Green-Refactor-Validate cycle. Use when implementing UI features, fixing UI bugs, or verifying fixes visually. Triggers include "TDD", "test-driven", "verify in browser", "spot check", "visual confirm", "red-green-refactor".

Execute Playwright E2E tests for QraftBox. Handles Nix environment setup, test execution, result directory management, and output analysis. Use when running, debugging, or analyzing E2E tests.

Use when implementation work in this repository changes behavior, adds functionality, or fixes bugs and the user has not explicitly asked to avoid workflows. Routes the work through the project-local divedra workflow `divedra-workflows/design-and-implement-review-loop`, including design/plan alignment, implementation, review, user-facing documentation refresh, and final handoff.

Use when creating or organizing design documents. Provides directory structure, file naming, and content guidelines for design specs and references.

Use when a divedra workflow, CLI command, GraphQL control plane, event listener, or submodule integration appears to fail because of an upstream divedra bug. Requires reproducing with the local submodule, filing an issue in tacogips/divedra when the bug is real, fixing the submodule, validating it, and committing the divedra change.

Use when helping an end user run, inspect, monitor, resume, continue, rerun, or troubleshoot existing divedra workflows. Applies to workflow list/status/validate/inspect/run, session progress/status/resume/continue/rerun, GraphQL inspection, mock scenarios, workflow roots, runtime artifacts, local serve, GraphQL endpoint execution, and event-triggered workflow usage.

Use when creating, modifying, reviewing, or validating divedra workflow bundles. Applies to step-addressed workflow.json authoring, node payload files, prompt files, built-in node add-ons, cross-workflow transitions, manager/worker routing, and portable workflows under a workflow root.

Use when testing or verifying divedra workflow bundles. Applies to deterministic mock-scenario runs, workflow validate/inspect/run checks, EXPECTED_RESULTS.md updates, regression assertions, session output verification, and CI-style workflow fixture validation.

Use when executing tasks from implementation plans. Provides task selection, parallel execution, progress tracking, and review cycle guidelines.

Use when creating implementation plans from design documents. Provides plan structure, status tracking, and progress logging guidelines.

Execute kinko release operations end-to-end for Go binaries, including local verification, artifact packaging, and optional GitHub release publishing. Use when users ask to release, publish a version, or build release binaries.

Use when users want to manage encrypted environment variables with the kinko CLI, including init/unlock, shared vs repo scope, set/get/show/delete/move, stale path-scope pruning, export/import, and exec-based runtime injection.

Use when creating or organizing design documents. Provides directory structure, file naming, and content guidelines for design specs and references.

Use when writing, reviewing, or refactoring Go code. Provides error handling patterns, project layout, concurrency, and interface design guidelines.

Use when creating implementation plans from design documents. Provides plan structure, status tracking, and progress logging guidelines.

Use this skill when creating or modifying GitHub Actions workflow files (.github/workflows/*.yml). Ensures all actions are pinned by commit SHA, permissions are minimized, script injection is prevented, and other supply chain security best practices are applied.

Use when writing Go code that interacts with dependencies, handles credentials, executes commands, or manages configuration. Provides supply chain attack countermeasures at the code level including safe dependency usage, credential handling, subprocess hardening, and runtime integrity patterns. Adapted from Shai-Hulud npm attack lessons.

Use when adding, updating, or auditing Go module dependencies. Provides supply chain attack countermeasures including go.sum verification, GOPROXY hardening, govulncheck auditing, and CI/CD pipeline security. Adapted from Shai-Hulud npm attack lessons.

Use when writing to X/Twitter with x-gateway-writer. Covers auth preflight, schema inspection, and project-owned GraphQL mutations such as createPost, quotePost, repostPost, unrepostPost, replyToPost, deletePost, likes, Lists, and DMs. Do not use x-gateway-reader for mutations.

Use when reading data from X/Twitter with x-gateway-reader. Covers auth preflight, schema inspection, and read-only project-owned GraphQL queries such as account lookup, post lookup, search, timelines, Lists, DMs, social graph reads, and usage. Do not use for posting or other mutations.

Build macOS Swift binary archives for the x-gateway reader and writer CLI executables.

Publish x-gateway artifacts to GitHub Releases using gh CLI with tag/version validation and idempotent retry behavior.

Use when building, validating, publishing, or tap-rendering Homebrew formula tarball releases for this Swift project, including scripts/build-homebrew-release.sh, scripts/render-homebrew-formula.sh, and task build:homebrew, homebrew:formula-reader, or homebrew:formula-writer commands.

Execute Nix release validation for x-gateway by verifying flake reproducibility, lockfile integrity, and optional lock updates.

Orchestrate x-gateway release operations end-to-end across Swift binary artifacts, GitHub release assets, Homebrew formulae, and Nix validation. Use when users ask to release or publish a version.

Use when resolving x-gateway issues through the packaged codex-design-and-implement-review-loop workflow, including accepted design and plan alignment, implementation, review, documentation refresh, verification, commit generation, and push.

Use when running live credential-backed smoke tests for rielflow Discord Gateway, Telegram Gateway, Matrix, or multi-persona chat workflows through Computer Use. Applies to opening Discord/Telegram/Element or Matrix web UIs, sending natural chat messages or attachments, checking persona routing, default responders, handoffs, history-after-restart behavior, and collecting rielflow event/session/reply evidence without exposing credentials.

Use when implementation work in a rielflow project changes behavior, adds functionality, or fixes bugs and the user has not explicitly asked to avoid workflows. Routes the work through the packaged `codex-design-and-implement-review-loop` workflow, including design/plan alignment, implementation, review, user-facing documentation refresh, commit-message generation, and built-in git commit/push steps.

Use when creating, modifying, reviewing, or validating rielflow workflow bundles. Applies to step-addressed workflow.json authoring, node payload files, prompt files, built-in node add-ons, cross-workflow transitions, manager/worker routing, and portable workflows under a workflow root.

Use when writing TypeScript code that interacts with dependencies, handles credentials, executes child processes, or manages configuration. Provides Shai-Hulud supply chain attack countermeasures at the code level including safe dependency usage, credential handling, subprocess hardening, and runtime integrity patterns.

Use when writing, reviewing, or refactoring TypeScript code. Provides type safety patterns, error handling, project layout, and async programming guidelines.

Use when the user asks Codex to security-check current source code, run secure code review, find and fix source security issues, run deterministic SAST or dependency/security scans, or loop fixes until security findings are resolved through the packaged `codex-source-security-check-loop` Rielflow workflow.

Use when preparing, publishing, verifying, or troubleshooting a rielflow release through GitHub release assets and the tacogips/tap Homebrew formula. Applies to building standalone Homebrew archives, rendering Formula/rielflow.rb, updating the sibling homebrew-tap repository, brew smoke tests, release rollback checks, and post-release install verification.

Use when bumping rielflow workspace package versions without publishing a release. Applies to requests such as bump to a requested version, version bump only, prepare version metadata, or updating rielflow package.json and bun.lock versions. Stops before tags, GitHub releases, archive builds, Homebrew tap updates, npm publishing, commits, or pushes unless the user explicitly asks for those actions.

Use when running or documenting the codex-design-and-implement-review-loop for Riela issue-resolution work packages, including accepted design, implementation plan, implementation, review, documentation refresh, verification, and final commit handoff.

Verify RielaApp macOS UI changes through the current debug executable, screenshots, AppKit layout tests, and stale-bundle checks. Use when reviewing, fixing, or visually validating RielaApp windows such as Instances, workflow viewer, prompts, profile selection, Settings-style lists, iPhone-like flows, or any UI spacing/alignment issue.

Create or improve deterministic Riela scenario tests that cover package creation, validation, packing, CLI/package workflow use, RielaApp import/readiness/enablement behavior, workflow execution monitoring, and adversarial Riela-based assessment.

Improve Riela and RielaApp usability through realistic user workflows. Use when Codex is asked to improve, redesign, harden, or review Riela/RielaApp features, specs, package flows, profile switching, workflow/package import, CLI help, onboarding, or any behavior where user friction should be found by operating the product rather than only reading code.

Verify Riela chat examples and live chat memory regressions through RielaApp, especially Telegram, Discord, Matrix, persona trio chats, image attachment prompts, raw-log/daily-summary memory examples, event-source receipts, and chat reply visibility checks.

Use when building, signing, notarizing, validating, publishing, or tap-rendering riela macOS Homebrew Cask `.dmg` artifacts.

Use when Codex is asked to implement, refactor, review, or maintain Swift code in SwiftPM, Xcode, iOS, macOS, watchOS, tvOS, or server-side Swift projects. This skill enforces SwiftLint execution, splits Swift files over 1000 lines into meaningful responsibility-based files, and guides maintainable Swift design using DRY, single responsibility, SOLID, established design patterns, tests, and project-local conventions.

Use when setting up or verifying Apple Developer ID signing credentials, Apple app-specific passwords, kinko secret storage, or local macOS notarization readiness for riela without recording credential values.

Use the installed apple-gateway macOS CLI to inspect or modify Apple Calendar events, Reminders, Notes, Mail, notifications, and Clock alarms; diagnose Apple app permissions; print the GraphQL schema; download materialized files; or prune the gateway cache. Use when Codex needs to work with the user's Apple app data through apple-gateway on Darwin.

Validate files that would be newly added to Git before commit or push, and detect content that should not be published such as secrets, private keys, local-only artifacts, machine-specific paths, private repository URLs, internal data, and generated junk. Use when a user asks whether new files are safe to commit or push.

Use when creating, publishing, or maintaining npm packages with Bun. Provides Shai-Hulud supply chain attack countermeasures including npm token management, 2FA enforcement, provenance signing, trusted publishing via GitHub Actions, and pre-publish security checklists.

Explain and list Lazydocker shortcuts for this repository, distinguishing repo-local config from upstream default keybindings. Use when a user asks for a Lazydocker shortcut list, what a Lazydocker key does, or whether this repo customizes Lazydocker keys.

Explain and list this repository's Yazi keymaps, including repo-local overrides and the shipped Yazi defaults. Use when a user asks what a Yazi key does, wants a Yazi keymap inventory, or wants to know the grep/find/search bindings in this repo.

Maintain and explain this repository's nvf-based Neovim keymap conventions. Use when a user wants to add, move, remove, audit, or document Neovim keymaps, especially when keeping Helix-like Space mode and goto-style g-prefixed navigation consistent.

Audit Alt/Meta (M-*) keymap conflicts for this repository's tmux, Yazi, Neovim (nvf), Zed/zeditor, and Lazygit setup. Use when a user wants to know which files define Alt bindings, how to inspect them, or whether bindings collide across terminal, tmux, editor, and file-manager layers.