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 skills add jazzenchen/VibeAround --skill agent-collaborationgit clone --depth 1 https://github.com/jazzenchen/VibeAroundWrote 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/jazzenchen/vibearound/agent-collaboration)<a href="https://agentmods.dev/skills/jazzenchen/vibearound/agent-collaboration"><img src="https://agentmods.dev/badge/skills/jazzenchen/vibearound/agent-collaboration.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00039 | $0.01024 |
| Opus 5 | $0.00019 | $0.00512 |
| Sonnet 5 | $0.00008 | $0.00205 |
| Haiku 4.5 | $0.00004 | $0.00102 |
Grade A, and why
agent-collaboration 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 — 121 lines — stays where its author put it; the contents beside it link to each section on GitHub.
VibeAround Agent Collaboration
Initialize a VibeAround multi-agent turn from the current host agent, then wait for subagent reports before finalizing.
When to Use
- The user message starts with
subagent= - The first supported mode is
subagent=parallel
Protocol
All host-to-subagent and subagent-to-host control messages must use va-agent-protocol.
VibeAround intercepts protocol envelopes in the thread. The raw envelope is hidden from the web UI and forwarded internally to the target agent. The protocol envelope must be the final content in the assistant message; do not write prose after it.
Use this envelope for structured messages:
<va-agent-protocol>
{
"protocol": "va-agent-protocol",
"kind": "assignment",
"turn_id": "<multi_agent_turn_id>",
"to_agent_id": "<subagent_guid>",
"task": "<clear task for this subagent>",
"context": "<only the relevant context>"
}
</va-agent-protocol>
Subagents must report back with:
<va-agent-protocol>
{
"protocol": "va-agent-protocol",
"kind": "report",
"turn_id": "<multi_agent_turn_id>",
"from_agent_id": "<subagent_guid>",
"status": "completed",
"summary": "<what changed or what was found>",
"files_changed": [],
"tests": [],
"notes": []
}
</va-agent-protocol>
Initialize Parallel Subagents
Call the VibeAround MCP tool:
Tool: va_mcp_initialize_subagents
Server: vibearound
Arguments:
thread_id: "<value of $VIBEAROUND_THREAD_ID>"
cwd: "<current working directory>"
mode: "parallel"
agents:
- name: "<host-chosen display name, e.g. John Planner>"
agent_kind: "codex"
task: "<parallel task>"
- name: "<host-chosen display name, e.g. Maya Reviewer>"
agent_kind: "codex"
task: "<parallel task>"
Then wait for the turn to finish:
Tool: va_mcp_wait_for_subagents
Server: vibearound
Arguments:
thread_id: "<value of $VIBEAROUND_THREAD_ID>"
turn_id: "<turn id returned by va_mcp_initialize_subagents>"
Rules:
- Choose concise human names for subagents. Names are display aliases; the MCP tool returns GUID agent IDs.
- For
parallel, default to exactly 2 subagents. - Use
codexfor subagents by default, even when the host is Claude Code. - Only create more than 2 subagents or use another
agent_kindwhen the user explicitly asks. - For
parallel, split the user's request into independent tasks that can run in separate git worktrees. - VibeAround handles missing Git repositories by running
git initand creating an empty initial commit when needed. If Git itself is missing, VibeAround attempts a platform install before reporting an error. - Do not merge or clean up worktrees automatically. The host agent decides after reviewing results.
- If VibeAround reports a dirty workspace or worktree creation error, tell the user and stop the multi-agent turn.
- VibeAround injects subagent role/system guidance at session startup. Keep assignments focused on the task and relevant context.
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 Changed dc392f19c733
- 9d ago First seen · 121 lines · 39 tokens per session scan A 327f140fd4a6
agent-collaboration is a skill published in the GitHub repository jazzenchen/VibeAround (523 stars, last pushed 3d ago), licensed MIT. It adds 39 tokens to every session and 1,024 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
opencli-sitemap-author
Use when creating or maintaining OpenCLI site sitemaps: agent-facing navigation, page-state, action, workflow, API-reference, pitfall, and fallback knowledge for a website. Use after browser exploration discovers durable site context, when a sitemap is stale, or when promoting local site knowledge into the repo.
omh-rust
This is a Hermes-native rust workflow skill.
omh-code-review
This is a Hermes-native code-review workflow skill.
redteam-web-detail-pack
Routing and boundary guidance for authorized general web application security testing. Use as a web testing router when the attack surface should be dispatched to more specific web vulnerability skills.
android-pentest
A guide for authorized security testing of Android apps, covering APK inspection, runtime testing, traffic capture, code review, and function hooking. An APK is the installable package used by an Android app.
jikegongyuan-perspective
A writing guide based on GeekPark, a Chinese technology publication and entrepreneur community focused on AI, products, startups, and innovation.