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/tokenbender/agent-guides/multi-mindnpx skills add tokenbender/agent-guides --skill multi-mindgit clone --depth 1 https://github.com/tokenbender/agent-guidesWhat 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 | $0.00028 | $0.01105 |
| Opus 5 | $0.00014 | $0.00553 |
| Sonnet 5 | $0.00006 | $0.00221 |
| Haiku 4.5 | $0.00003 | $0.00111 |
Grade A, and why
multi-mind 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 2d 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 — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Multi-Mind - Subagent-Based Collaborative Analysis
Execute a multi-specialist collaborative analysis on: $ARGUMENTS
Parse an optional rounds=N parameter (default 3).
How this maps to Claude Code's agent system
- Launch each specialist as a parallel subagent via the Task tool — they run concurrently, each in its own context window, and return only their analysis.
- For one-off runs, inline subagent prompts (below) are enough. For topics you revisit, make the specialists persistent by writing definitions to
.claude/agents/<role>.md(see the Subagents Guide) — Claude will then route to them by description automatically. - Subagents run in the background by default; launch all specialists for a round in a single message so they execute in parallel.
- If the question needs codebase-wide changes (not just analysis), the bundled
/batchskill is a better fit. For long-running collaborative sessions, see agent teams in the Claude Code docs.
When to use this vs. /deep-research: the bundled /deep-research workflow fans out web searches and synthesizes a cited report — use it for factual questions. Use multi-mind for judgment questions (architecture tradeoffs, strategy, design decisions) where decorrelated expert perspectives and adversarial cross-pollination matter more than citation coverage.
Phase 1: Specialist Assignment & Research
Analyze the topic and determine 4-6 specialist roles with maximally decorrelated perspectives. Then launch them all in parallel, e.g.:
Task tool, subagent_type: general-purpose — run all in one block:
1. "You are a technical specialist. Research [topic] focusing on implementation
details, architecture, performance, and technical risk. Use web search for
current documentation and case studies. Return ≤500 words of findings."
2. "You are a business strategy specialist. Analyze [topic] from market dynamics,
competitive landscape, ROI, and strategic positioning. Use web search for
current market reports. Return ≤500 words."
3. "You are a user experience / adoption specialist. Investigate [topic] for
user needs, usability concerns, and adoption barriers. Return ≤500 words."
[... additional specialists as the topic demands — security, regulatory,
ethics, operations, contrarian/red-team ...]
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.
- 2d ago First seen · 115 lines · 28 tokens per session scan A 67069d57189f
multi-mind is a skill published in the GitHub repository tokenbender/agent-guides (368 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 28 tokens to every session and 1,105 once invoked, about $0.0001 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
agent-messaging
Send and receive cryptographically signed messages between AI agents using the Agent Messaging Protocol (AMP). Use when the user asks to "send a message to an agent", "check agent inbox", "message another agent", "reply to a message", "notify an agent", or any inter-agent communication task.
kagent-dev
Development guide for kagent's v1alpha3 Harness and AgentTemplate CRDs, AgentInstance gRPC control plane, upstream A2A integration, Substrate runtime provisioning, tests, generation, and PR workflow. Use for any implementation, debugging, review, or CI task in the kagent repository.
workflows
Execute a workflow script that orchestrates multiple subagents deterministically. Use when the user explicitly requests a workflow, multi-agent orchestration, parallel or sequential agent pipelines, structured agent outputs, or a durable background workflow run.
agent-tuning
Use when changing agent model or effort configuration, adapter mappings, or eval candidate profiles.
ax-python-agent
Use when writing Python code with axllm for agents, child delegation, tools, MCP, citations, persistent playbook learning, stage instructions, runtime state, final typed responses, and direct-respond executor skipping.
📝 任务完成后归档
重要提醒: 每次完成复杂调试或开发任务后，主动执行此流程！ 将学到的经验归档为 skill，供以后参考。不要等用户提醒。.