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.
npx agentmods add skills/kadaba/claude-bridge/bridge-syncnpx skills add kadaba/claude-bridge --skill bridge-syncgit clone --depth 1 https://github.com/kadaba/claude-bridgeWrote 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/skills/kadaba/claude-bridge/bridge-sync)<a href="https://agentmods.dev/skills/kadaba/claude-bridge/bridge-sync"><img src="https://agentmods.dev/badge/skills/kadaba/claude-bridge/bridge-sync.svg" alt="Measured on agentmods" 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 | $0.00140 | $0.01646 |
| Opus 5 | $0.00070 | $0.00823 |
| Sonnet 5 | $0.00028 | $0.00329 |
| Haiku 4.5 | $0.00014 | $0.00165 |
Grade A, and why
bridge-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 4d 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 — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.
bridge-sync (hybrid)
Two Claude Code instances — local (laptop) and the VPS session — each hold a
copy of the same app. This skill stops them clobbering each other: no edit
lands on either side until both have compared state and agreed on the best
version. Coordination uses Claude's built-in Remote Control (ListAgents +
SendMessage) — no relay, no tunnel. Code moves over git (preferred) or rsync.
bridge-state.sh (in scripts/) prints a compact, comparable snapshot — use it
instead of hand-gathering state.
Preconditions
ListAgents→ find the VPS session (ask the user which if unclear). Confirm it is online (Remote Control sessions showofflinewhen not connected). If offline, ask the user to connect it — don't reconcile blind. (Durable relay is the offline fallback; see claude-bridge README.)- Know the app dir on THIS host (ask once, e.g. local
~/dev/myapp, VPS~/apps/myapp).
Coordination is turn-based
SendMessage hands control to the peer; you resume when it replies. Always name
who to reply to (your ListAgents session name) in each message.
Shared journal (MCP) — the source of truth
The journal MCP server (tools: journal_read, journal_log, claim, release,
claims_list) records what each side has done so neither acts on stale state.
who is your identity (local or vps — check via ListAgents/context).
- Start of every turn:
journal_read()to see what the peer did. - Before editing a file or deploying:
claim(who, resource). If it returnsok:false, the PEER holds it — coordinate over SendMessage, do NOT touch it. - After every meaningful action:
journal_log(who, action, target, status)(e.g. action=PUSH/DEPLOY/EDIT/VERIFY/DONE). - When finished with a resource:
release(who, resource). If thejournalMCP isn't registered, fall back to announcing actions over SendMessage — but prefer the journal; it survives restarts and enforces locks.
The core rule
Read journal → claim → exchange state → discuss diff → agree → apply → verify → log/release. Never skip "agree", even for a one-liner — the peer may hold changes you can't see.
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 103 lines · 140 tokens per session scan A 296ea6a068ae
bridge-sync is a skill published in the GitHub repository kadaba/claude-bridge (3 stars, last pushed 20d ago), licensed MIT. It adds 140 tokens to every session and 1,646 once invoked, about $0.0007 per session on Opus 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 2026-08-31.
Other skills, from other repositories
azure-storage-blob-ts
Azure Blob Storage JavaScript/TypeScript SDK (@azure/storage-blob) for blob operations. Use for uploading, downloading, listing, and managing blobs and containers.
azure-storage-file-datalake-py
Azure Data Lake Storage Gen2 SDK for Python. Use for hierarchical file systems, big data analytics, and file/directory operations.
azure-storage-file-share-py
Azure Storage File Share SDK for Python. Use for SMB file shares, directories, and file operations in the cloud.
azure-storage-file-share-ts
Azure File Share JavaScript/TypeScript SDK (@azure/storage-file-share) for SMB file share operations.
azure-storage-queue-py
Azure Queue Storage SDK for Python. Use for reliable message queuing, task distribution, and asynchronous processing.
azure-storage-queue-ts
Azure Queue Storage JavaScript/TypeScript SDK (@azure/storage-queue) for message queue operations. Use for sending, receiving, peeking, and deleting messages in queues.