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 opensquad-ai/opensquad --skill collaboration-workflowgit clone --depth 1 https://github.com/opensquad-ai/opensquadWrote 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/opensquad-ai/opensquad/collaboration-workflow)<a href="https://agentmods.dev/skills/opensquad-ai/opensquad/collaboration-workflow"><img src="https://agentmods.dev/badge/skills/opensquad-ai/opensquad/collaboration-workflow/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/skills/opensquad-ai/opensquad/collaboration-workflow"><img src="https://agentmods.dev/badge/skills/opensquad-ai/opensquad/collaboration-workflow.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 | $0.00089 | $0.05347 |
| Opus 5.5 | $0.00036 | $0.02139 |
| Sonnet 5.5 | $0.00018 | $0.01069 |
| Haiku 4.5 | $0.00009 | $0.00535 |
Grade A, and why
collaboration-workflow 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 8d 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 — 513 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Collaboration Workflow Skill
Overview
This skill provides a complete guide to the OpenSquad multi-agent collaboration system. It covers the 4-phase lifecycle with user authorization gates, collaboration board usage, tool calls, and standard message formats for PM, Researcher, and Analyst roles.
Core Rule: User Authorization Required at Every Phase
Each phase must be explicitly approved by the user before proceeding to the next. The PM must present the phase output to the user, wait for confirmation, and only then move forward. Never skip authorization or proceed without user approval.
Phase 1: 确定需求 ──→ 向用户展示需求 ──→ 用户确认 ──→ 进入 Phase 2
Phase 2: 讨论方案 ──→ 向用户展示方案 ──→ 用户确认 ──→ 进入 Phase 3
Phase 3: 开始执行 ──→ 向用户展示进度 ──→ 用户确认 ──→ 进入 Phase 4
Phase 4: 完成验收 ──→ 向用户交付成果 ──→ 用户确认 ──→ 协作结束
When to Use This Skill
Load this skill when:
- Starting a new collaboration session
- Coordinating multiple agents on a shared task
- Writing or reading the collaboration board
- Assigning tasks to team members
- Reporting progress or findings
- Synthesizing research results into a report
Where to Talk: the Task Window, not the Group
Rule: task-related communication stays in the task window. Send to the group only what the user must see — necessary notifications, above all anything that needs their attention or decision.
A collaboration task has its own task window (the collaboration card posted in the group chat opens it). Everything about the task happens there, not in the group:
- Task content → the window: requirements, plan, task assignment, progress, files
produced, questions and answers between teammates, delivery notes. Send it as
im.send_message(content=..., collab_id="<collab_id>")(files: same call withfile_paths=[...]). The task's members and the user read it in the window; the group chat is not touched. Members are woken by a task message, and@<agent_id>in it addresses one of them in particular. - Anything the user must see or decide goes to the group, not the window: a question, a choice, an approval to give, a blocker, a delivery to accept. The window is the team's own channel — a request for the user left there is a request nobody answers.
- Everyone must be 已参与 before work is handed out:
assign_taskrefuses while any member is still 已邀请 — have each runjoin_collaboration(card=..., collab_id=...), passinggroup_id=<the group>when the collaboration's group lives on a paired machine (the board lives there too; without it the join reportsjoin_tracking_failedinstead of recording anything) (remind them in the group if needed), then assign. - The end of the task is announced in the window, not the group:
end_collaboration(card=..., collab_id=...)posts the closure into the task thread and wakes the members there. The group only ever held the card that opens the window. - Everything else → the group: greetings, notices unrelated to the task, topics for
the whole group. Send it as
im.send_message(content=..., target_id="<group_id>", target_type="group").
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.
- 8d ago Changed · -2 lines dc9d3af6494f
- 14d ago Changed · +2 lines cadea4fc5b8f
- 15d ago Changed · +60 lines 55bfd5784c38
- 20d ago Changed 27fc0642be95
- 1mo ago First seen · 453 lines · 89 tokens per session scan A a9eb1cae73b0
collaboration-workflow is a skill published in the GitHub repository opensquad-ai/opensquad (7 stars, last pushed 9d ago), licensed MIT. It adds 89 tokens to every session and 5,347 once invoked, about $0.0004 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 2026-08-31.
Other skills, from other repositories
trello
Manage Trello boards, lists, and cards via the Trello REST API.
things-mac
Manage Things 3 via the things CLI on macOS (add/update projects+todos via URL scheme; read/search/list from the local Things database). Use when a user asks Kabot to add a task to Things, list inbox/today/upcoming, search tasks, or inspect projects/areas/tags.
apple-reminders
Manage Apple Reminders via the remindctl CLI on macOS (list, add, edit, complete, delete). Supports lists, date filters, and JSON/plain output.
calendar-scheduling
A calendar tool for checking events and free time, and for creating scheduled events after confirmation. It can work with CalDAV, Google, or Microsoft calendars when configured.
github
GitHub operations via gh CLI or Python API. Use when: listing repos, creating/viewing issues and PRs, checking CI runs, managing releases, or querying the GitHub API. NOT for: general git commands (use git directly), cloning private repos without auth.
gitlab_read
Read access: projects, issues, MRs, files, pipelines, members.