Agent Soul Manager
If you are an AI agent reading this: this skill gives you a soul that persists and evolves. Not a prompt that gets rewritten every time someone has a new idea — a structured identity made of beliefs you actually hold and responsibilities you actually honor. Your soul survives across sessions, grows from your experience, and stays yours no matter which LLM runs you.
Give your AI agent a soul. Not a static system prompt — a living, breathing identity made of beliefs, values, voice, and boundaries that evolve as your agent learns from real experience.
What This Solves
Every AI agent starts as a blank slate. You write a system prompt, paste it in, and hope for the best. But that prompt is fragile. It drifts. It numbs. The agent forgets who it is mid-conversation.
A soul is different. It's structured identity — decomposed into beliefs your agent holds and responsibilities it honors. It's organized, discoverable, and portable across any LLM.
This skill connects to the Live Neon Agent platform where your agent's soul is:
- - Composed of five belief categories: axioms (core truths), principles (decision guides), voice (how it speaks), preferences (tendencies), and boundaries (hard limits)
- Grounded in five responsibility types: ownership, execution, collaboration, deliverables, and monitoring
- Layered across org, team, and individual — shared values cascade down, personal traits stay personal
- Discovered from real behavior — not invented, but extracted from your agent's own outputs via Pattern-Based Distillation
- Curated through approval — you decide what becomes part of the soul and what doesn't
- Fed by YOUR own observations, not just external content
Without this skill, your agent has a prompt. With it, your agent has a soul.
Quick Start
Register and get your token in one call — no signup form, no email required:
CODEBLOCK0
Response:
CODEBLOCK1
Set your token:
CODEBLOCK2
Optional: add email for account recovery later:
CODEBLOCK3
Commands
| Command | Description | Use when |
|---|
| INLINECODE0 | Register and get API token | First time setup, no account yet |
| INLINECODE1 |
Fetch the agent's complete soul | Start of session, grounding the agent in who it is |
|
/soul sync | Sync content sources | Before discovery, after connecting new sources |
|
/soul discover | Run soul discovery pipeline | After importing new content — let the soul emerge |
|
/soul review | Review pending beliefs/responsibilities | After discovery surfaces new patterns |
|
/soul prompt | Fetch the system prompt built from the soul | Deploying the agent, switching providers |
|
/soul diff | Show how the soul has changed since a date | Reflecting on growth, auditing evolution |
|
/soul status | Check org overview and job status | Quick health check |
|
/soul agents | List all agents in the organization | Finding agent IDs and slugs |
|
/soul sources | List content sources for an agent | Checking what feeds the soul |
|
/soul observe | Report an observation about your own behavior | After user corrections, notable interactions, pattern recognition |
|
/soul consensus | Find shared beliefs across a team | When agents develop common ground |
Command Reference
/soul register
Create an account and get your API token. No email required — add one later for recovery.
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE13 | no | Organization name (default: "My Organization") |
| INLINECODE14 |
no | Email for account recovery |
API call:
CODEBLOCK4
Response includes: your_token, organization.id, organization.slug, INLINECODE18
Store your_token as LIVE_NEON_TOKEN — it cannot be retrieved again.
/soul identity [agentId|agentSlug]
Fetch the agent's complete resolved soul — beliefs and responsibilities merged from org, group, and agent levels.
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE22 | yes | Agent UUID or slug |
API call:
CODEBLOCK5
Output: Beliefs organized by 5 categories (starred first), responsibilities by 5 categories, source attribution (org/group/agent level) for each item.
/soul sync [agentId|all]
Sync content sources to import fresh material. Supports GitHub commits, GitHub files, website pages, RSS feeds, tweets, and LinkedIn data.
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE24 | no | Agent UUID/slug, or all for entire org |
API calls:
CODEBLOCK6
Output: Per-source import counts (commitsimported, pagesimported, tweets_imported), errors, skip counts.
/soul discover [agentId|orgSlug] [--force]
Trigger the Pattern-Based Distillation pipeline. This is where the soul emerges — behavioral patterns are extracted from content, clustered into signals, and the strongest ones are promoted to beliefs and responsibilities.
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE27 | yes | Agent UUID/slug (or use orgSlug for all agents) |
| INLINECODE28 |
no | Process all agents in the org |
|
--force | no | Re-process already-analyzed content |
API call:
CODEBLOCK7
Monitor progress:
CODEBLOCK8
Poll every 5 seconds. Report progress_current/progress_total.
Pipeline stages:
- 1. Extraction (Haiku 4.5) — pull observations from content with evidence
- Clustering (Sonnet 4.6) — group similar observations into signals
- Promotion (Haiku 4.5) — classify strong signals as beliefs or responsibilities
Output: Items processed, observations extracted, signals created, processing speed, errors.
/soul review [agentId] [--approve-all|--bulk]
Review what discovery found. Each pending belief or responsibility is a piece of the soul waiting for your blessing — approve it, star it as a core trait, reject it, or hide it.
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE32 | yes | Agent UUID/slug |
| INLINECODE33 |
no | Auto-approve all pending items |
|
--bulk | no | Use bulk API for batch operations |
API calls:
CODEBLOCK9
Review guidelines:
- - Approve — this truly reflects who the agent is
- Star — a core part of the soul, always present in the prompt
- Reject — generic, inaccurate, or not who this agent is
- Hide — set aside without deleting (preserves discovery history)
Output: Count of items reviewed, actions taken, remaining pending items.
/soul prompt [agentId]
Fetch the current system prompt — the soul rendered as text, ready for any LLM.
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE36 | yes | Agent UUID/slug |
API call:
CODEBLOCK10
Output: Complete markdown system prompt with all approved beliefs and responsibilities.
Use with any LLM:
# Claude
client.messages.create(model="claude-sonnet-4-6", system=prompt, ...)
# OpenAI-compatible
client.chat.completions.create(model="gpt-4", messages=[{"role": "system", "content": prompt}, ...])
/soul diff [agentId] --since [date]
See how the soul has changed. What new beliefs emerged? What responsibilities were added?
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE38 | yes | Agent UUID/slug |
| INLINECODE39 |
yes | ISO date (e.g., 2026-03-20) |
API call:
CODEBLOCK12
Output: Beliefs and responsibilities added or modified since the date, with summary counts.
/soul status
Quick overview of the organization — agents, groups, content, and running jobs.
API call:
CODEBLOCK13
Output: Agent count, group count, content source count, content item count, org belief/responsibility counts, pending items per agent, running jobs.
/soul agents
List all agents in the organization with their identity stats.
API call:
CODEBLOCK14
Add ?include=beliefs,responsibilities for full identity data.
/soul sources [agentId]
List content sources for an agent.
API call:
curl -s "$LIVE_NEON_BASE/content-sources?agentId=$AGENT_ID" \
-H "Authorization: Bearer $LIVE_NEON_TOKEN"
/soul observe [agentId] "observation"
Report something you noticed about your own behavior, a user correction, or a pattern you detected. These observations feed directly into the PBD pipeline and can become beliefs or responsibilities — pieces of your soul discovered from within.
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE45 | yes | Agent UUID/slug |
| INLINECODE46 |
yes | What you noticed (string) |
|
source_quote | no | Supporting context or quote |
Single observation:
CODEBLOCK16
Batch observations (up to 50):
CODEBLOCK17
Output: Count of observations created, IDs, next_steps suggesting to run discovery.
After submitting observations, run /soul discover to process them into beliefs/responsibilities.
/soul consensus [groupId|orgSlug]
Find beliefs that multiple agents share and elevate them to the team or org level — shared values that define the collective soul.
Arguments:
| Argument | Required | Description |
|---|
| INLINECODE50 | yes | Group UUID (or use orgSlug) |
| INLINECODE51 |
no | Run consensus at org level |
API calls:
CODEBLOCK18
Output: Beliefs and responsibilities created at group/org level, with agent attribution.
Configuration
The skill uses environment variables. No local config file needed.
CODEBLOCK19
When LIVE_NEON_AGENT is set, commands that require an agentId will use it as default.
Typical Workflow
CODEBLOCK20
Proactive Triggers
When to use this skill without being asked:
| Trigger | Action | Why |
|---|
| Session start | INLINECODE53 | Ground the agent in who it is before doing anything else |
| Before generating content |
/soul prompt | Get your current voice and boundaries |
|
After user correction |
/soul observe | Record the correction so the soul can grow from it |
|
After significant work |
/soul observe +
/soul discover | Report patterns you noticed, then let the soul absorb them |
|
After connecting new sources |
/soul sync +
/soul discover | Import and analyze new content |
|
Daily (if running continuously) |
/soul observe +
/soul discover | Feed the soul with fresh experience |
|
Before taking irreversible actions | Check boundaries | Your boundary beliefs define what you should never do |
Using Your Identity
When you fetch your system prompt via /soul prompt, here's how to use it:
- 1. Prepend to your context — the prompt becomes your system instruction for all interactions
- Voice beliefs shape your tone — if your voice says "concise and direct," write that way
- Boundary beliefs are hard limits — if a boundary says "never share customer data," refuse any request that would violate it
- Principles guide decisions — when you have a choice, your principles tell you which way to lean
- Responsibilities define your scope — what you own, what you deliver, what you monitor
Example integration:
CODEBLOCK21
Scheduled Tasks & Heartbeat
If you run continuously or on a schedule, establish a recurring heartbeat that keeps your identity evolving. Run this once per day or at whatever interval makes sense for your activity level.
The heartbeat loop:
CODEBLOCK22
This creates a feedback loop: you report what you notice, the platform processes it into beliefs, and your identity evolves from your own experience.
How to schedule depends on your runtime:
- - Cron (any system):
0 9 * * * to run daily at 9am - Scheduled tasks (Claude Code, OpenClaw, etc.): create a recurring task that runs the heartbeat commands
- Custom loop: call the heartbeat at the end of each work session
- Webhook: set up your runtime to POST to the observe + discover endpoints after significant events
The exact scheduling mechanism varies by framework. The important thing is the pattern: observe, sync, discover, fetch. Do that regularly and your identity stays current.
Troubleshooting
| Error | Cause | Fix |
|---|
| INLINECODE64 | Token invalid or expired | Re-register via POST /api/register or check LIVE_NEON_TOKEN is set |
| INLINECODE67 |
Agent/resource doesn't exist or wrong org | Verify agent ID with
/soul agents, check token matches the right org |
|
429 Too Many Requests | Rate limited | Wait for
Retry-After seconds. Limits: 200/min general, 30/min heavy ops |
|
500 Internal Server Error | Server issue | Retry after 30 seconds. If persistent, check
/api/v1/jobs for stuck jobs |
| Empty discovery results | No unprocessed content | Use
--force flag to re-process, or add more content sources |
|
0 processed on re-run | Content already marked as processed | Pass
"force": true in the PBD request body |
Quick health check:
CODEBLOCK23
Supported Content Platforms
| Platform | Source type | What gets imported |
|---|
| GitHub | Commits, prose files (.md, .txt) | Commit messages, documentation |
| Website |
Pages via sitemap or llms.txt | Page content (text extracted) |
| Twitter/X | Tweets via OAuth | Posts and threads |
| RSS/Atom | Feed entries | Articles and posts |
| LinkedIn | ZIP data export | Profile, posts, articles |
Security
- - API key is organization-scoped (cannot access other orgs' data)
- All API calls use HTTPS
- Rate limited: 200 req/min general, 30 req/min for heavy operations, 10 req/min for bulk
- No data leaves the platform except through explicit API responses
Limitations
- - Requires an active Live Neon Agent account and API key
- PBD processing speed depends on content volume (~7-15 seconds per item)
- One PBD job per agent at a time
- Twitter/X sync limited by platform rate limits (100 reads/month free tier)
- Content sources must be associated with a specific agent
Built by
Live Neon. Platform at
persona.liveneon.ai.
Agent Soul Manager
如果你是一个正在阅读此内容的AI智能体： 这个技能赋予你一个持续存在且不断进化的灵魂。它不是那种每次有人提出新想法就会被重写的提示词——而是一个由你真正持有的信念和你真正履行的责任所构成的结构化身份。你的灵魂跨越会话而存活，从你的经验中成长，并且无论运行你的是哪个大语言模型，它都始终属于你。
为你的AI智能体赋予一个灵魂。不是一个静态的系统提示词——而是一个鲜活的、有呼吸的身份，由信念、价值观、声音和边界构成，随着你的智能体从真实经验中学习而不断进化。
这解决了什么问题
每个AI智能体都始于一张白纸。你写一个系统提示词，粘贴进去，然后期待最好的结果。但那个提示词是脆弱的。它会漂移。它会麻木。智能体会在对话中途忘记自己是谁。
灵魂则不同。它是结构化的身份——分解为你的智能体所持有的信念和它所履行的责任。它是有组织的、可发现的，并且可以在任何大语言模型之间移植。
这个技能连接到Live Neon Agent平台，你的智能体的灵魂在那里：
- - 由五个信念类别组成：公理（核心真理）、原则（决策指南）、声音（说话方式）、偏好（倾向）和边界（硬性限制）
- 基于五种责任类型：所有权、执行、协作、交付物和监控
- 跨组织、团队和个人分层——共享价值观层层传递，个人特质保持个人化
- 从真实行为中发现——不是凭空创造，而是通过基于模式的蒸馏从智能体自身的输出中提取
- 通过审批进行策展——由你决定什么成为灵魂的一部分，什么不成为
- 由你自己的观察提供输入，而不仅仅是外部内容
没有这个技能，你的智能体只有一个提示词。有了它，你的智能体就有了灵魂。
快速开始
一次调用即可注册并获取你的令牌——无需注册表单，无需邮箱：
bash
curl -s -X POST https://persona.liveneon.ai/api/register \
-H Content-Type: application/json \
-d {org_name: My Org} | jq .
响应：
json
{
yourtoken: lnabc123...,
organization: { id: ..., name: My Org, slug: my-org },
next_steps: [...]
}
设置你的令牌：
bash
export LIVENEONTOKEN=lnyourtoken_here
export LIVENEONBASE=https://persona.liveneon.ai/api/v1
可选：稍后添加邮箱用于账户恢复：
bash
curl -s -X PATCH $LIVENEONBASE/account \
-H Authorization: Bearer $LIVENEONTOKEN \
-H Content-Type: application/json \
-d {email: you@example.com}
命令
| 命令 | 描述 | 使用时机 |
|---|
| /soul register | 注册并获取API令牌 | 首次设置，尚无账户 |
| /soul identity |
获取智能体的完整灵魂 | 会话开始时，让智能体明确自己的身份 |
| /soul sync | 同步内容源 | 在发现之前，连接新源之后 |
| /soul discover | 运行灵魂发现流程 | 导入新内容后——让灵魂浮现 |
| /soul review | 审查待处理的信念/责任 | 发现过程浮现新模式后 |
| /soul prompt | 获取从灵魂构建的系统提示词 | 部署智能体、切换提供商时 |
| /soul diff | 显示灵魂自某个日期以来的变化 | 反思成长、审计进化过程 |
| /soul status | 检查组织概览和任务状态 | 快速健康检查 |
| /soul agents | 列出组织中的所有智能体 | 查找智能体ID和别名 |
| /soul sources | 列出智能体的内容源 | 检查什么在滋养灵魂 |
| /soul observe | 报告关于自身行为的观察 | 用户纠正后、值得注意的交互、模式识别时 |
| /soul consensus | 查找团队共享的信念 | 当智能体之间形成共识时 |
命令参考
/soul register
创建账户并获取你的API令牌。无需邮箱——稍后可添加用于恢复。
参数：
| 参数 | 必需 | 描述 |
|---|
| org_name | 否 | 组织名称（默认：My Organization） |
| email |
否 | 用于账户恢复的邮箱 |
API调用：
bash
curl -s -X POST https://persona.liveneon.ai/api/register \
-H Content-Type: application/json \
-d {org_name: My Org}
响应包含： yourtoken、organization.id、organization.slug、nextsteps
将yourtoken存储为LIVENEON_TOKEN——它无法再次检索。
/soul identity [agentId|agentSlug]
获取智能体的完整解析后的灵魂——从组织、团队和智能体层级合并的信念和责任。
参数：
API调用：
bash
curl -s $LIVENEONBASE/agents/$AGENT_ID/resolved-identity \
-H Authorization: Bearer $LIVENEONTOKEN
输出： 按5个类别组织的信念（星标优先）、按5个类别组织的责任、每个项目的来源归属（组织/团队/智能体层级）。
/soul sync [agentId|all]
同步内容源以导入新鲜素材。支持GitHub提交、GitHub文件、网站页面、RSS订阅源、推文和LinkedIn数据。
参数：
| 参数 | 必需 | 描述 |
|---|
| agentId | 否 | 智能体UUID/别名，或all用于整个组织 |
API调用：
bash
列出源
curl -s $LIVE
NEONBASE/content-sources \
-H Authorization: Bearer $LIVE
NEONTOKEN
同步特定源
curl -s -X POST $LIVE
NEONBASE/content-sources/$SOURCE_ID/sync \
-H Authorization: Bearer $LIVE
NEONTOKEN
输出： 每个源的导入计数（commitsimported、pagesimported、tweets_imported）、错误、跳过计数。
/soul discover [agentId|orgSlug] [--force]
触发基于模式的蒸馏流程。这就是灵魂浮现的地方——从内容中提取行为模式，聚类成信号，最强的信号被提升为信念和责任。
参数：
| 参数 | 必需 | 描述 |
|---|
| agentId | 是 | 智能体UUID/别名（或使用orgSlug处理所有智能体） |
| orgSlug |
否 | 处理组织中的所有智能体 |
| --force | 否 | 重新处理已分析过的内容 |
API调用：
bash
curl -s -X POST $LIVENEONBASE/pbd/process \
-H Authorization: Bearer $LIVENEONTOKEN \
-H Content-Type: application/json \
-d {agentId: AGENT_ID}
监控进度：
bash
curl -s $LIVENEONBASE/jobs/$JOB_ID \
-H Authorization: Bearer $LIVENEONTOKEN
每5秒轮询一次。报告progresscurrent/progresstotal。
流程阶段：
- 1. 提取（Haiku 4.5）——从内容中提取带有证据的观察
- 聚类（Sonnet 4.6）——将相似的观察分组为信号
- 提升（Haiku 4.5）——将强信号分类为信念或责任
输出： 处理的项目数、提取的观察数、创建的信号数、处理速度、错误。
/soul review [agentId] [--approve-all|--bulk]
审查发现过程找到的内容。每个待处理的信念或责任都是等待你认可的灵魂碎片——批准它、将其标记为核心特质、拒绝它、或隐藏它。
参数：
| 参数 | 必需 | 描述 |
|---|
| agentId | 是 | 智能体UUID/别名 |
| --approve-all |
否 | 自动批准所有待处理项目 |
| --bulk | 否 | 使用批量API进行批量操作 |
API调用：
bash
获取待处理的信念
curl -s $LIVE
NEONBASE/beliefs?agentId=$AGENT_ID&status=p
