DevAgentRadar → Google Antigravity

Goal: turn this release signal into a safe pilot plan for my stack.

Context

Assistant: Google Antigravity Proposal: Hide model-only skills with disable-slash-command and script print-mode slash output Summary: Antigravity 1.1.12 lets teams declutter / via SKILL.md disable-slash-command: true, script read-only slash commands with -p (TSV/JSON, no agent turn), and capture clean models/agents lists via --output-format json. Primary source: https://github.com/google-antigravity/antigravity-cli/releases/tag/1.1.12

Why it matters

Large skill libraries flood the / menu even when many skills are only for the model. 1.1.12 adds disable-slash-command: true in SKILL.md frontmatter: the skill stays discoverable and invocable by the model but is hidden from the / menu and /name resolution. Print mode now answers the read-only slash commands /permissions, /hooks, /help, /changelog, and /config with one tab-separated record per line — or a structured payload under --output-format json and stream-json — without starting an agent turn, spending quota, or leaving a conversation. /help lists exactly the commands print mode answers; /changelog prints the release notes. The models and agents subcommands accept the same --output-format values, and their errors and progress spinner moved off stdout so captured output is only the list. Put the frontmatter on model-only skills and switch inventory scripts to -p plus --output-format so menus stay usable and automation never burns a turn.

Suggested actions

1. Find every SKILL.md that is meant for the model only (not for users to type). Add disable-slash-command: true to its YAML frontmatter so it stays discoverable and invocable by the model but is hidden from the / menu and from /name resolution.
2. Run print mode for each newly non-interactive slash command as TSV and as structured output: -p "/permissions"; -p "/hooks"; -p "/help"; -p "/changelog"; -p "/config"; then the same five with --output-format json and again with --output-format stream-json. Confirm each emits records without starting an agent turn, spending quota, or leaving a conversation, and that -p "/help" lists exactly those print-mode commands.
3. Change every script that lists models or agents to pass --output-format json (or stream-json). Treat stdout as the list only — errors and the progress spinner are no longer on stdout — and fail if the payload is not valid JSON.
4. Replace scraped changelog and config/permission/hook checks with -p "/changelog", -p "/config", -p "/permissions", and -p "/hooks" (add --output-format json where the consumer is a parser) so those audits never start a turn or spend quota.

Config surfaces this release may change

• background and headless runs (high confidence) — check your repo before applying
• skills — check your repo before applying

After you finish

Do not report this as applied to DevAgentRadar. You cannot write the visitor's loop.

Tell the human: open https://devagentradar.com/proposals/google-antigravity-1-1-12-hide-model-only-skills-with-disable-slash-comm and mark Applied, Skipped, or Failed. Proposal id: 1a2003d7-8abb-4163-b4ad-8ce278ca756b

Please:

1. Map the change to concrete pilot steps
2. Flag security / permission implications
3. Keep the pilot reversible