gpambrozio

baseline-review

Review E2E baseline image changes in the current PR.

Testing & QA Listed

Automatically create E2E test scenarios that prove a new feature works. Use this skill when a PR or branch introduces a new feature and you want to generate an end-to-end scenario that exercises it, takes screenshots at key points, and verifies the screenshots show the expected behavior. Trigger when: the user says "create e2e for this feature", "write an e2e test for this PR", "prove this feature works with e2e", "add e2e coverage for this change", or any variation of wanting automated E2E proof that new code works. Also use when the user asks to "test this feature end to end" or "create a scenario for the new feature". Do NOT use for fixing existing failing scenarios (use e2e-testing), manual debugging (use e2e-manual-debugging), or reviewing baselines (use baseline-review).

e2e-manual-debugging

Use this skill when an e2e scenario can't find a UI element — `iosTap`/`iosWaitForElement`/`macClickButton`/`macWaitForElement` keeps timing out, you don't know the right `ElementQuery` to use, you need to know what `AXLabel` / `AXIdentifier` / `.help()` / `accessibilityLabel` a button actually exposes, or `iosLogUI` output isn't enough. This skill boots an interactive e2e-mode instance of the apps and walks through inspecting the running UI (XCUITest view hierarchy, AppleScript accessibility tree, Xcode's Accessibility Inspector, on-the-fly screenshots) to discover the exact attributes a scenario step needs. Use whenever someone says "I can't find this element", "what's the right query for X", "the test can't see Y", "what label does the macOS button expose", "find the accessibility identifier", or "the button isn't being clicked". Do NOT use this skill for writing new scenarios from scratch (use e2e-for-feature) or running/fixing scenarios when the issue isn't an unknown element (use e2e-testing).

Testing & QA Listed

e2e-testing

Use this skill when writing, modifying, running, or debugging formal e2e test scenarios — the Swift DSL-based automated tests in ClaudeSpyE2ELib. This includes creating new scenario files, adding or changing TestStep sequences, fixing failing scenarios, choosing ElementQuery matchers, registering scenarios in allScenarios, adding accessibility identifiers/hooks for testability, running ./scripts/e2e-test.sh, updating screenshot baselines, or extending the test framework with new step types. Use this skill whenever someone mentions e2e scenarios, test steps, element queries, screenshot baselines, the ScenarioBuilder DSL, or wants to run the automated e2e suite. Do NOT use this skill for ad-hoc manual debugging, taking one-off screenshots, interactively driving the app, or exploratory testing without writing a formal scenario — use e2e-manual-debugging for those instead.

Testing & QA Listed

fix-e2e-failures

Investigate and fix E2E test failures from the latest test report. Use this skill when the user mentions E2E failures, broken tests, test report errors, failing screenshots, baseline mismatches, updating baselines, or wants to fix a failing E2E scenario. Also use when someone says "fix the test", "deal with the test failure", "check the e2e report", "update baselines", or references a PR comment about E2E failures. Do NOT use this for writing new E2E tests (use e2e-testing) or for manual debugging (use e2e-manual-debugging).

install

Install app on one or multiple targets

AI & Automation Listed

create-agent-plugin

Build a new Gallager sidecar plugin from scratch — a standalone executable that teaches the Gallager macOS app to monitor a coding agent (track session state, raise attention badges, fire notifications, push text into panes). Use this skill whenever someone wants to add support for a coding agent that Gallager doesn't already handle (anything beyond the built-in Claude Code and Codex), wire an agent's hooks into Gallager, write or debug a "sidecar plugin", build a Gallager plugin, or distribute one via URL. Trigger on phrases like "make a Gallager plugin for <agent>", "I want Gallager to track my <CLI/agent> sessions", "add <agent> support to Gallager", "write a sidecar plugin", "create an agent plugin", or "how do I get my agent's hooks into Gallager".

gallager-cli

Control the Gallager macOS app from the command line to manage tmux sessions, windows, panes, send text/keys to terminals, trigger desktop notifications, and open files in the in-app prompt editor. Use this skill whenever the user is working inside a Gallager-managed tmux session and wants to drive the UI from scripts or the shell — creating sessions, switching panes, sending input, notifying when a long-running task finishes, or editing prompts from the current terminal. Trigger on phrases like "open a new tmux window", "send this to the pane", "notify me when build finishes", "split pane", "edit this prompt in Gallager", or any mention of the `gallager` CLI.