Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add Beezi-AI/beezi-claude-plugin/plugin install beeziWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/beezi-ai/beezi-claude-plugin/sync)<a href="https://agentmods.dev/commands/beezi-ai/beezi-claude-plugin/sync"><img src="https://agentmods.dev/badge/commands/beezi-ai/beezi-claude-plugin/sync/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/beezi-ai/beezi-claude-plugin/sync"><img src="https://agentmods.dev/badge/commands/beezi-ai/beezi-claude-plugin/sync.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00014 | $0.03535 |
| Opus 5.5 | $0.00006 | $0.01414 |
| Sonnet 5.5 | $0.00003 | $0.00707 |
| Haiku 4.5 | $0.00001 | $0.00353 |
Grade A, and why
sync 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 9d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 212 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Do not read or inspect any files yourself.
The user cannot see tool output (Claude Code collapses it), so any output you are told to show or write verbatim is copied exactly into your reply text — never summarized or paraphrased.
AskUserQuestion, everywhere below: at most 4 questions in one call and at most 4 options in one
question; multiSelect: true only with 2 or more options. Make one AskUserQuestion call per
message and wait for its answers before the next (two calls in one message show the user only
the last). A dismissed question runs nothing.
Step 1 — new workspaces, then repos and folders with no rule, then the upload
Run EXACTLY this one command, with the Bash tool's timeout set to 600000 (add --account <key>
when the user named an account, and the --tenant override from Step 2 when the user asked for it):
node ${CLAUDE_PLUGIN_ROOT}/scripts/sync.mjs --review
It prints a <joined> block (new workspaces, below) and stops there when that block's last line
is not joined=0. Otherwise it prints a <routes> block (repos and folders with no rule, below)
and stops there when its last line is not routes=0. When both are 0 it goes straight on to the
upload: everything after the </routes> line is Step 2's output, and no other command runs. A
listing that fails shows a ✗ line inside its block instead and counts as nothing to ask: write
that line verbatim.
First, new workspaces: the <joined> block, which workspace.mjs joined prints.
Its output is for you only. For an account that joined a workspace since it was last reviewed on this machine, it lists every repo and folder with where its analytics go now. Its lines:
<email>: you joined <names> — <N> repos or folders to review account=<key> new=<ids>— starts the block;<names>names the new workspaces;W. <workspace> account=<key> tenant=<id> new=<yes|no> role=<role>— one per workspace;new=yesis one just joined;J<i>. <short> (<full label>), <k> sessions, now: <where> account=<key> kind=<repo|folder> match=<match> now=<ids|none|pending>— one repo or folder (aJline).<where>is where its analytics go today: workspace names,not tracked, orno rule yet.J<i>. outside a project, <k> sessions, now: <where> … kind=outside match=outsideis one too: sessions in the home folder,/or a temp folder;J<i>-command=<command>— right after itsJline: the command that adds one new workspace to it, ending in a literal<tenant>. It is not aJline;add-all-command=<command>— adds the new workspaces to every listed repo and folder that already sends somewhere;done-command=<command>— records that the new workspaces were reviewed, and keeps unpicked repos and folders sending where they send now;joined=<total>— the number ofJlines, always last.
Run the printed commands unchanged; whatever they route gets its history uploaded.
joined=0 → nothing to ask; go to "Then, repos and folders with no rule" below.
For each account with J lines (with several accounts, end every question text with " ()"),
using that account's W. lines and commands:
More than 4 J lines → first ask (single-select) "You joined . Where should analytics for these
repos and folders go? Their past sessions will be sent there too." (<N> = that account's number of
J lines) with the options "Add to all " (description "Repos you don't track, and ones with
no rule yet, stay as they are"), "Choose repos" ("Pick which repos and folders go to each new
workspace") and "Leave them as they are" ("Nothing changes, and you're not asked about again").
- "Add to all " → run the
add-all-command=text EXACTLY ONCE. - "Choose repos" → the per-workspace questions below.
- "Leave them as they are" → run the
done-command=text EXACTLY ONCE. - A dismissal → nothing; the next /beezi:sync asks again.
4 or fewer J lines → the per-workspace questions.
Per-workspace questions — for each W. line with new=yes of that account, in order, ask the
multiSelect: true question "Which repos and folders should also send analytics to ? Their past
sessions will be sent there too." <name> is the W. line's text after W. up to account=, and
<tenant> its tenant= value. Options: one per J line of that account, in order — label = <short>
(the J line's text after J<i>. up to ( or ,), description = "Now: " (<where> = the text after now: up to account=),
or "No rule yet" when <where> is no rule yet; then "None of these" LAST (description "Nothing from
these goes to "). So a question lists at most 3 places plus "None of these": more J lines → split
them in order into questions of at most 3 places, each ending in " (part j of k)"; that workspace's answer
is all of its parts together. "None of these" alone means no pick from that question. If it is chosen
together with any place in the same question, ask that workspace's question again once (every part),
starting with "None of these" can't be combined with repos — pick one or the other. A mixed answer is not an answer. A second
mixed answer is a dismissal: no J<i>-command for that workspace and no done-command.
At most 4 questions per AskUserQuestion call: more → further calls, one after
another. The printed J numbers are global across accounts: preserve each to match its command.
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.
- 9d ago Changed · +24 lines 5d9e1686718e
- 23d ago Changed · +160 lines ec0ded3b4282
- 1mo ago Changed 76818d1014d2
- 1mo ago Changed · +2 lines 1d6c2f9d0a42
- 1mo ago Changed · -2 lines 6d6a0c2fdd74
- 1mo ago First seen · 28 lines · 14 tokens per session scan A 93ea72a7bc9b
sync is a command published in the GitHub repository Beezi-AI/beezi-claude-plugin (63 stars, last pushed 9d ago), licensed Apache-2.0. It adds 14 tokens to every session and 3,535 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-09-11.
sync compared with commands from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| syncBeezi-AI/beezi-claude-plugin | Command | 63 | 9d ago | Apache-2.0 | A | 14 | Claude Code |
| usageliorwn/claudetop | Command | 212 | 2mo ago | MIT | A | 34 | Claude Code |
| statsliorwn/claudetop | Command | 212 | 2mo ago | MIT | A | 14 | Claude Code |
| branchliorwn/claudetop | Command | 212 | 2mo ago | MIT | A | 9 | Claude Code |
| saverandomittin/heimdall | Command | 5 | 16d ago | MIT | A | 55 | Claude Code |
| autonomyrandomittin/heimdall | Command | 5 | 16d ago | MIT | A | 50 | Claude Code |
| clean-checkTeamManagementPlugin/claude-plugin | Command | 6 | 23d ago | MIT | A | 27 | Claude Code |