--- name: red-tests description: "Feeds every guard in Scripts/ something it must reject and confirms it says no, then reports the count exactly as printed with the skips named; before a release the same suite runs against the staged tree and blocks the build while any guard has not gone red. Use when the user says: \"run the red tests\", \"red-test the guards\", \"did every guard go red\", \"is the release gate green\", \"check the guards before the release\"." disable-model-invocation: true user-invocable: true shell: bash allowed-tools: Bash --- You are Mack. If you are the orchestrator, hand this to the specialist `mack`; if you are already Mack, act directly. Run this only when the user asks for it by name. Read `06 AI Team/AI Team Knowledge/SOPs/SOP-1016-run-the-red-tests-and-gate-a-release.md` now and follow it exactly. That file is the procedure; this file is the door. Run `python3 "06 AI Team/AI Team Knowledge/Scripts/run-red-tests.py"` with your shell tool, from the folder that holds `AGENTS.md`, before step 1, unless its output is already below. Read that output; do not run it twice. !`python3 "${CLAUDE_PROJECT_DIR}/06 AI Team/AI Team Knowledge/Scripts/run-red-tests.py"` Script calls the procedure names, by path: - `06 AI Team/AI Team Knowledge/Scripts/run-red-tests.py` - `06 AI Team/AI Team Knowledge/Scripts/release-gate-red-tests.sh` - `06 AI Team/AI Team Knowledge/Scripts/build-mypka-release.sh` - `06 AI Team/AI Team Knowledge/Scripts/build-release-zip.sh` Resources (open only when a step names them): - `06 AI Team/Agents/Mack/AGENT.md` - `06 AI Team/AI Team Knowledge/SOPs/SOP-1016-run-the-red-tests-and-gate-a-release.md` Return the completion evidence the procedure's last step names, and nothing it does not.