Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
/plugin marketplace add ArshiaEcho/janitornpx agentmods add plugins/arshiaecho/janitor/marketplacegit clone --depth 1 https://github.com/ArshiaEcho/janitorGrade A, and why
janitor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
{
"name": "janitor",
"owner": {
"name": "Stratos House AI",
"url": "https://www.stratosagency.ai"
},
"metadata": {
"description": "A Claude Code marketplace shipping the janitor plugin: two skills for cleaning up slow machines (macOS and Windows) and messy Claude Code working directories. Reversible, permission-gated, no telemetry.",
"homepage": "https://www.stratosagency.ai"
},
"plugins": [
{
"name": "janitor",
"description": "Two slash-command skills: janitor-device diagnoses and cleans up slow macOS or Windows machines, and janitor-claude audits and tidies the current Claude Code working directory. Read-only by default, every mutation requires explicit confirmation, no silent sudo, no git mutations, no telemetry.",
"category": "productivity",
"source": "./plugins/janitor",
"homepage": "https://github.com/ArshiaEcho/janitor",
"keywords": [
"cleanup",
"performance",
"mac",
"macos",
"windows",
"claude-code",
"maintenance",
"diagnostics",
"obsidian",
"vault"
]
}
]
}
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 33 lines scan A cada062410c3
janitor is a plugin published in the GitHub repository ArshiaEcho/janitor (6 stars, last pushed 2mo ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other plugins, from other repositories
hal-9000
Plugin marketplace listing 3 plugins: hal-voice, hal-session-auto-rename, hal-skills.
apple-mail-fast-mcp
Apple Mail for Claude Code — search, read, organize, draft, templates, rules, and inbox stats via MCP, with an IMAP fast path. macOS only.
apple-pim
Native macOS integration for Calendar, Reminders, Contacts, and Mail using EventKit, Contacts, and JXA frameworks.
imessage-mcp
macOS-only MCP server that exposes iMessage history to Claude and exports per-chat conversations into an Obsidian vault. Direct chat.db read, Whisper voice-note transcription, FTS5 search, AppleScript send with draft+confirm.
claude-speak
Plugin marketplace listing 1 plugin: claude-speak.
relay
Delegate work packets from a lead Claude Code session to executor sessions in their own iTerm tabs, with session reuse instead of always spawning fresh.