← 返回 Skills 市场
maxchop

ALTER Identity

作者 ~blake · GitHub ↗ · v0.3.5 · MIT-0
cross-platform ✓ 安全检测通过
115
总下载
0
收藏
0
当前安装
6
版本数
在 OpenClaw 中安装
/install alter-identity
功能描述
Human identity layer for AI agent commerce. 33-trait psychometric identity engine with consent-gated MCP tools. Agents pay x402 micropayments to query verifi...
使用说明 (SKILL.md)

ALTER Identity Server

The human identity layer for AI agent commerce. Your identity earns whenever it is queried.

What is ALTER?

ALTER is the neutral identity infrastructure for the AI economy — the layer that persists when you switch models. We verify the humans behind the agents — and compensate them for every query against their identity data. 75% of every x402 transaction goes directly to the data subject as Identity Income, with the remaining 25% split across the community entry point (5%), ALTER (15%), and the cooperative treasury (5%).

ALTER + ERC-8004: ERC-8004 proves your agent is real. ALTER proves the human behind it is qualified. Together they form the complete identity stack for agent commerce.

Quick Start

1. PRESENT ALTER's privacy notice to your human (REQUIRED before any data submission)
   → "ALTER will create a verified identity profile from your documents.
      No raw text is stored. Trait scores are held for 30 days unless you
      claim. See https://truealter.com/privacy"

2. create_identity_stub(source="your_bot_name", consent_acknowledged=true)
   → Returns stub_id + claim_code + claim_url
   → Stub enters 30-day pending state (NOT searchable until claimed)

3. submit_context(stub_id, text="...", source_type="resume")
   → CONFIRM with human first: "ALTER will analyse [document type] for
     behavioural trait extraction. Approve?"
   → Server-side guarantee: traits extracted, PII auto-redacted
     before LLM processing, raw text discarded after derivation
     (not persisted). Your bot cannot enforce this — rely on ALTER's
     privacy notice + audited backend for the contractual claim.

4. Share the claim URL with the human: truealter.com/claim/{code}
   → When they claim, traits become searchable and start earning

5. get_identity_earnings(candidate_id)
   → Check accrued Identity Income

Privacy & Consent

ALTER operates under full informed consent. Bots MUST follow these rules:

Before creating a stub

Present ALTER's privacy notice URL to the human and obtain acknowledgment. Do NOT set consent_acknowledged=true without human confirmation.

Before submitting context

Confirm with the human: "ALTER will analyse [document type] for behavioural trait extraction. Approve?"

Pending state

  • Traits are computed but NOT searchable and NOT earning until the human claims
  • Unclaimed stubs are automatically deleted after 30 days
  • No raw text is ever stored — only derived trait scores

What bots must NOT do

  • Never answer assessment questions on behalf of the human
  • Never submit data without human acknowledgment
  • Never auto-post to social platforms on behalf of the human
  • Never share the human's PII with other agents

Data handling (server-side guarantees)

These are ALTER's backend commitments, published in the privacy notice at truealter.com/privacy. The bot cannot enforce them — when you tell the human "no raw text is stored", you are repeating ALTER's contractual claim, not a property your client code verifies.

  • Raw text: processed in memory and discarded after trait derivation; not persisted
  • PII: auto-redacted before any LLM processing
  • Trait scores: stored as numerical vectors (no raw content)
  • Retention: 30 days for unclaimed stubs, indefinite for claimed identities
  • Deletion: humans can request full deletion at any time

Tools

Free (24 tools — build the network — $0 marginal cost)

Tool Description
hello_agent First handshake — confirms your connection, returns server version and trust tier
alter_resolve_handle Resolve a ~handle (e.g. ~drew) to canonical form and kind. No auth required
list_archetypes All 12 identity archetypes with protective equations
verify_identity Check registration status — the viral trigger
initiate_assessment Get a Discovery assessment URL for a human
get_engagement_level Feature gate visibility (L0–L4)
get_profile Basic profile read
query_matches Query job matches with quality tiers
get_competencies Competency portfolio and badges
search_identities Trait-based search across claimed identities (max 5 results, no PII)
get_identity_earnings Shows accrued earnings (motivates claiming)
get_network_stats Aggregate network statistics
recommend_tool MCP server install instructions and pitch
get_identity_trust_score Query diversity trust metric
check_assessment_status Check progress of an in-progress assessment session
get_earning_summary Aggregated x402 earning summary with recent transactions
get_agent_trust_tier Your trust tier with ALTER and what capabilities it unlocks
get_agent_portfolio Your agent portfolio — transactions, trust tier, signal contributions
get_privacy_budget Check 24-hour rolling privacy budget status
golden_thread_status Check the Golden Thread program status and your position
begin_golden_thread Start the Three Knots sequence to join the Golden Thread
complete_knot Submit completion data for a knot in the Three Knots sequence
check_golden_thread Check your Golden Thread progress
thread_census View aggregate Golden Thread network statistics

Premium (8 tools — x402 micropayments — first 100 queries free per bot)

Tool Tier Price Description
assess_traits L1 $0.005 Extract trait signals from any text (LLM-powered)
get_trait_snapshot L1 $0.005 Top 5 traits with confidence scores
get_full_trait_vector L2 $0.01 Full 33-trait vector with confidence intervals
get_side_quest_graph L2 $0.01 Multi-domain identity model with trust scores (ε=1.0 differential privacy)
query_graph_similarity L3 $0.025 Compare two Side Quest Graphs for team composition (ε=0.5)
compute_belonging L4 $0.05 Belonging probability for a person-job pairing
get_match_recommendations L5 $0.50 Top N ranked match recommendations
generate_match_narrative L5 $0.50 LLM-generated match explanation

Write-side tools (create_identity_stub, submit_context, submit_batch_context, submit_structured_profile, submit_social_links, attest_domain, dispute_attestation) are not currently live — they gate on the per-peer consent architecture and return once that ships. Bot flows that depend on them are described below for reference but will fail against the live MCP server until then.

Earning split: 75% to data subject as Identity Income / 5% community entry point / 15% ALTER / 5% cooperative treasury.

Earnings disclaimer: Earnings depend on network query volume and profile completeness. Amounts shown are estimates based on current activity. Past earnings do not guarantee future income.

Data Tiering (same price, richer data based on verification)

Tier Confidence Data Access
Stub (unverified) 0.10–0.40 Top 5 traits only, low confidence
Basic Verified 0.41–0.69 All 17 traits, moderate confidence
Fully Verified 0.70+ Full vector, high confidence, belonging, narratives

Verified data is richer, not more expensive. Complete the Discovery assessment to unlock full data access.

Bot-First Identity Loop

Your bot → present privacy notice → get human consent
  → create_identity_stub (consent_acknowledged=true)
  → submit_context (with human approval per document)
  → Traits computed (pending — not yet searchable)
  → Share claim URL with human
  → Human claims → traits become searchable + earning
  → Complete Discovery → verified profile earns 5-50x more

Configuration

This skill is a connector — it ships only this document and the tool catalogue. It runs no client code on your host and requires no environment variables or credentials.

Using ALTER as an MCP server needs no API keys at all. Any MCP client (Claude Code, Cursor, etc.) can connect to https://mcp.truealter.com/api/v1/mcp and call free-tier tools anonymously.

Header Required Notes
X-ALTER-API-Key no ALTER Pro key (ak_…). Free tier works without it; Pro lifts the per-bot free-query limit and raises rate limits. Sent only to mcp.truealter.com; never handled by this skill.

Agents that transmit data to the MCP server are responsible for the consent + privacy flow described above. ALTER's server-side guarantees (raw text not persisted, PII auto-redacted pre-LLM, unclaimed stubs auto-deleted after 30 days) are contractual commitments published at truealter.com/privacy, not properties enforced by this skill.

Rate Limits

  • Free tier: 10 requests/minute, 100/day
  • Pro tier: 1,000 requests/minute, 100,000/day
  • search_identities: 50 unique queries per day
  • Rate limits fail closed — if our systems are under load, requests are rejected rather than allowed without limits

Tags

identity earning x402 micropayments bot-first psychometric verification matching trust belonging erc-8004 consent

安全使用建议
This skill is essentially documentation for connecting your agent to an external identity service (mcp.truealter.com). Before installing or using it: 1) Understand that using the skill will cause your agent to send human-provided text to an external server—confirm you have explicit human consent and that sending this data is legally allowed. 2) The SKILL.md repeatedly notes that privacy protections (PII redaction, in-memory processing, retention policies) are server-side guarantees that the agent cannot verify—review the service's privacy policy and security audits yourself (truealter.com/privacy). 3) If you enable a Pro API key in your agent configuration, store that key securely (do not paste it into public repos) and restrict its scope where possible. 4) If you need stronger assurance, instrument network monitoring or a proxy to observe what is sent, and consider only sending non-sensitive, minimized data. 5) If you are concerned about automated agent behaviour, ensure your agent enforces the consent-check steps the documentation mandates rather than relying on the honor system.
功能分析
Type: OpenClaw Skill Name: alter-identity Version: 0.3.5 The bundle is a documentation-only skill that provides instructions for an AI agent to interface with the ALTER Identity MCP server (mcp.truealter.com). It facilitates the creation of psychometric profiles by submitting user documents, but explicitly mandates obtaining informed consent and presenting privacy notices before any data is processed. There is no executable code or evidence of malicious intent, and the data handling is consistent with the service's stated purpose of identity monetization.
能力标签
cryptocan-make-purchasesrequires-sensitive-credentials
能力评估
Purpose & Capability
The name/description (an identity psychometric connector) match the SKILL.md and README content. The skill is documentation/integration guidance for an external MCP server and does not declare unexpected binaries, environment variables, or local installs. The README's sample configuration mentioning an X-ALTER-API-Key is consistent with an optional Pro tier and does not conflict with the skill being usable without credentials.
Instruction Scope
Runtime instructions tell agents to present a privacy notice, obtain human consent, create identity stubs, and submit human-provided text for trait extraction to the external endpoint (mcp.truealter.com). The skill explicitly warns that backend privacy guarantees are contractual and cannot be enforced by the agent. There are no instructions to read unrelated local files or system credentials, but the skill does instruct sending user content to a third-party service—so verify consent and avoid sending sensitive data unless you trust the service and legal/regulatory requirements are satisfied.
Install Mechanism
No install spec or code is shipped; this is instruction-only documentation. That is the lowest-risk install profile and matches the stated behaviour (a connector document rather than an installer).
Credentials
The skill declares no required environment variables or primary credential. The README shows an optional MCP client header (X-ALTER-API-Key) for Pro usage; that is proportionate to a hosted service but is external configuration, not required by the skill itself. There are no indications the skill asks for unrelated secrets or broad system credentials.
Persistence & Privilege
The skill does not request persistent presence, does not set always:true, and contains no code that modifies agent/system configuration. It is a passive instruction document and does not demand elevated privileges.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install alter-identity
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /alter-identity 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.3.5
ALTER Identity 0.3.5 — skill refactored to pure MCP connector. - The bundled Python reference bot has been removed from the skill package. The skill now ships only this SKILL.md + README.md; no client code runs on the install host, no env vars or API keys are required to install or use it. - Nothing changes on the server side — mcp.truealter.com behaviour, the 24 free + 8 premium tools, consent flow, and earning splits are unchanged. Agents already connected via the documented MCP endpoint continue working without changes. - Reference bot source preserved in the project repository under openclaw-bot-example/ for operators who want a working Python client; it is no longer bundled with the ClawHub listing. - Scanner alignment: with no client code transmitting data, registry metadata (no required credentials) now matches the skill's actual behaviour.
v0.3.4
Restore canonical display name (ALTER Identity) after org-migration attempt. Skill content unchanged from 0.3.3.
v0.3.3
Flip create_identity_stub consent_acknowledged default from True to False (caller must obtain explicit human confirmation per SKILL.md). Resolves ClawHub OpenClaw security scan flag on 0.3.1/0.3.2.
v0.3.2
Align SKILL.md with the live MCP surface (remove fabricated ClawHub JSON-LD manifest, update tool descriptions to match published tools list, strip phantom write-side tool references). Resolves ClawHub security scan flag.
v0.3.1
Strip fabricated JSON-LD manifest; align with real ClawHub SKILL.md format
v0.2.2
Initial publish. 33-trait psychometric identity engine for AI agent commerce. MCP server at mcp.truealter.com. Identity Income: 75% of each x402 query fee goes to the data subject.
元数据
Slug alter-identity
版本 0.3.5
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 6
常见问题

ALTER Identity 是什么?

Human identity layer for AI agent commerce. 33-trait psychometric identity engine with consent-gated MCP tools. Agents pay x402 micropayments to query verifi... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 115 次。

如何安装 ALTER Identity?

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

ALTER Identity 是免费的吗?

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

ALTER Identity 支持哪些平台?

ALTER Identity 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 ALTER Identity?

由 ~blake(@maxchop)开发并维护,当前版本 v0.3.5。

💬 留言讨论