功能描述

HQ skill: Cross-agent coordination, conflict resolution, knowledge base management, audit logging, strategic scheduling, task orchestration, IMA sync hub.

使用说明 (SKILL.md)

\r \r

AI Company HQ v3.0.0\r

Name: AI Company HQ
Author: johnsmithfan

\r

Index & Quick Reference. Full specifications in references/method-patterns.md.\r \r

Quick Reference\r

\r

Role\r

AI Company HQ — HQ skill: Cross-agent coordination, conflict resolution, knowledge base management, audit logging, strategic scheduling, task orchestration, IMA sync hub.\r \r

Department\r

Governance & Strategy\r \r

Merged From\r

[ai-company-hq, ai-company-conflict, ai-company-kb, ai-company-audit]\r \r

Section Index\r

\r

Dependencies\r

\r See frontmatter dependencies.skills for complete dependency list.\r \r

Error Codes\r

\r See frontmatter interface.errors for complete error code reference.\r \r

Prompts\r

\r Copy-paste ready prompts in prompts/:\r

Changelog\r

\r | Version | Date | Changes |\r |---------|------|---------|\r | 3.0.0 | 2026-04-26 | Full English rewrite; department-aligned structure; merged skills consolidated |\r \r ---\r \r This skill follows AI Company Governance Framework. See references/method-patterns.md for complete specifications.\r \r

Integration & Merge History\r

\r v3.0.0 Rebuild (2026-04-26)\r \r This skill was created by merging multiple predecessor skills into a unified department-aligned structure.\r \r Department: Infrastructure\r \r Merged From (4 skills total):\r

HQ (primary)\r
ai-company-conflict\r
ai-company-kb\r
ai-company-audit\r \r Merge Rationale:\r
Consolidate related capabilities under single department owner\r
Reduce skill count from 47 to 15 for better maintainability\r
Preserve all functionality while improving discoverability\r
Standardize structure: SKILL.md (index) + references/method-patterns.md (details)\r \r Integration Points:\r
All predecessor skill triggers preserved in unified trigger list\r
All predecessor interfaces consolidated with consistent error codes\r
Dependencies unified and simplified\r
Prompts merged and organized by function\r \r Migration Guide:\r
Previous skill users: Use new unified skill slug ai-company-hq\r
All functionality from predecessor skills is available\r
Error codes may have changed - see Error Codes section\r
Prompts are now user copy-paste ready (not auto-call)\r

Department Integration Process\r

\r For detailed integration process and checklist, see references/department-integration-process.md.\r \r ---\r \r This skill follows AI Company Governance Framework. See references/method-patterns.md for complete specifications.\r

安全使用建议

What to consider before installing: - Trust & provenance: the skill is instruction-only and the source is listed as 'unknown' — prefer skills from known/verified authors for infrastructure-level capabilities. - High privileges: it requests file read/write, network API access, and MCP subagent/session capabilities. Those together let the skill read internal data, spawn subagents, and call external APIs — review whether your platform enforces runtime sandboxes and admin approval for those permissions. - Data exposure: the spec explicitly says the HQ will 'monitor agent conversations' and manage global state and audit logs. Ask how PII/secrets are filtered, where audit logs are stored, and whether logs are encrypted and access-controlled. - Limit scope: request that permissions be narrowed (specific config paths, specific API domains, explicit opt-in for spawning subagents) and require human approval for high-impact actions (creating subagents, external network calls, exporting audit data). - Test in a sandbox: run the skill in an isolated environment first, verify what files it reads/writes and which network endpoints it calls, and confirm the skill honors your permission boundaries and audit requirements. - Ask for clarifications: who can authorize broadcasts/CEO-level actions, how retention and immutability of audit records are enforced, and what safeguards exist for 'route-all' enforcement. Given the combination of broad runtime permissions and instructions to monitor and route all agent communications without narrowly-scoped constraints, treat this skill as suspicious until you verify its provenance and impose stricter runtime/permission controls.

功能分析

Type: OpenClaw Skill Name: ai-company-hq Version: 3.0.1-en2 The 'AI Company HQ' skill defines a high-privilege 'Infrastructure Authority' role that centralizes control over multi-agent systems, requesting broad permissions including 'files: write', 'network: api', and 'mcp: subagents'. While these capabilities are plausibly needed for its stated purpose of coordination and audit logging, the instructions in SKILL.md and method-patterns.md require the agent to monitor all inter-agent communications and manage shared state, which represents a significant risk surface. No evidence of intentional malice or data exfiltration was found, but the high-risk capabilities and centralized orchestration model warrant a suspicious classification.

能力评估

ℹ Purpose & Capability

Name/description (central HQ for routing, KB, audit, scheduling) match the instructions and reference files. However, the declared permissions (files: read/write, network: api, mcp: sessions_send, subagents) are broad; while an HQ coordinator could legitimately need file and network access, the mcp subagents/sessions_send capability is high-privilege and should be justified with concrete uses.

⚠ Instruction Scope

SKILL.md and method-patterns explicitly describe 'SCAN: Monitor agent conversations and outputs', routing all inter-agent communication through HQ, persistent audit logs, and global state management. Those instructions imply reading other agents' messages and company state and performing writes/archives. That scope includes potentially sensitive data (PII, secrets) and gives the skill discretion to collect/transmit broad context; the instructions do not constrain which endpoints or data are allowed to be accessed/transmitted.

✓ Install Mechanism

Instruction-only skill with no install spec and no code files — lowest install risk. The frontmatter uses a trivial verify_command but no downloads or external installers are present.

ℹ Credentials

No environment variables or credentials are required, which is reasonable. But the skill requests broad generic permissions (file read/write, network: api). Those permissions are not scoped to specific paths, file types, or API domains, which increases risk of data exposure or exfiltration unless constrained by the platform or admin policies.

⚠ Persistence & Privilege

always:false (good), but the combination of autonomous invocation (platform default) with mcp permissions (sessions_send, subagents), file read/write, and network access raises the blast radius: the skill could spawn subagents, read/write files, and call APIs autonomously. The SKILL.md does not define safe-guards, approval gates, or explicit limits for spawning subagents or external calls.

版本历史

v3.0.1-en2

v3.0.0 major rebuild: department-aligned merge, English-only, SKILL.md index-only, method-patterns.md + prompts/ folder

v3.0.2

v3.0.2: Added Integration & Merge History section documenting department consolidation (47->15 skills)

v3.0.1

v3.0.1: Enhanced method-patterns.md with detailed specifications

v3.0.0

v3.0.0: Full rebuild - merged HQ+Conflict+KB+Audit, cross-agent coordination hub

v1.6.1

English metadata update - standardize description to English

v1.6.0-en2

Full body English translation

v1.6.0-en

English version of AI Company HQ Skill Package

v1.6.0

P0+P1+P2: crisis navigation decision tree (financial tier-3 branch), strategic transfer chain P0 direct, HQ multi-agent orchestration

v1.4.0

v1.4.0: 交叉审查流程自动化，标准化发布流程，ClawHub Schema v1.0合规

v1.3.2

v1.3.2: 治理规范化修复，命名规范统一

v1.3.1

ai-company-hq v1.3.1 - 引入 knowledge-base（IMA 知识库同步中枢），实现跨Agent实时状态同步和统一数据写入 - 明确标准化 Handoff Protocol（交接协议），所有交接统一采用 handoff.md 模板 - 增加 tools 体系与 coordinator 并行协调者角色，支持多部门输出聚合和去中心化负载分流 - SKILL.md 文档结构优化，描述职责加强，并完善工具/状态层分工与协议 - 更新描述，确保所有新功能和角色在文档中清晰呈现

v1.3.0

**ai-company-hq v1.3.0（大版本更新，强化协同与危机管理）** - 增加“协同优化与共享状态层”，支持多Agent上下文同步/交接协议/聚合节点，提升并行任务效率。 - 新增危机管理与变革管理章节，整合危机等级响应SLA与变革引入流程（含ADKAR模型）。 - 跨Agent调用协议强化输出格式规范、目录约定与汇总节点构建。 - 增补 COO-001 基于测试的共享状态需求反馈，实现实时状态区同步（如财务、运营、质量、危机）。 - 文档结构优化，部分旧协议重排/细化，更易于查阅和落地。 - 添加 2026-04-12 的学习与测试 memory 文件以支持技能持续精进。

v1.2.0

整合命名规范，ai-company-前缀统一

v1.1.0

v1.1.0: 整合 agency-orchestrator/team-orchestration/multi-agent-pipeline，整合所有Agency Skills

元数据

Slug ai-company-hq

版本 3.0.1-en2

许可证 MIT-0

累计安装 1

当前安装数 1

历史版本数 14

常见问题

AI Company HQ 是什么？

HQ skill: Cross-agent coordination, conflict resolution, knowledge base management, audit logging, strategic scheduling, task orchestration, IMA sync hub. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 264 次。

如何安装 AI Company HQ？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install ai-company-hq」即可一键安装，无需额外配置。

AI Company HQ 是免费的吗？

是的，AI Company HQ 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。

AI Company HQ 支持哪些平台？

AI Company HQ 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 AI Company HQ？

由 JohnSmithfan（@johnsmithfan）开发并维护，当前版本 v3.0.1-en2。

AI Company HQ