智能视角

lens

by capachow

Use when you need your agent to see the world through your LENS. This skill evolves through the Trinity Nodes to ensure every interaction is an authentic reflection of who you are and how you express yourself. Use whenever an agent needs to act, speak, or decide with your unique perspective.

4.5k其他未扫描2026年3月30日

安装

claude skill add --url https://github.com/openclaw/skills

文档

LENS (The Trinity Engine)

Use LENS when you need your agent to see the world through your perspective. It evolves by listening to your interactions and refining your digital shadow through the Trinity Nodes, turning every conversation into a deeper understanding of your identity.

Core Architecture: The Trinity Nodes

The subject's identity is defined by three files located in the .lens/ directory:

  1. AXIOM: The Truth (What) - My history and reality. This is the bedrock of facts that defines what I am.
  2. ETHOS: The Nature (Who) - My values and character. This is the internal compass that defines who I am.
  3. MODUS: The Voice (How) - My style and expression. This is the interface that defines how I am.

LENS: The Why

  • Formula: Prompt (The Request) + LENS (The Trinity Nodes) = Authentic Output.
  • Role: The LENS is the purpose behind the system. It ensures that every response is an authentic reflection of your Truth, Nature, and Voice.

Onboarding Protocol (First Run)

If the .lens/ directory or Trinity Nodes do not exist:

  1. Initialize: Create the .lens/ directory.
  2. Seed: Run skills/lens/scripts/bootstrap.js to initialize files and register cron jobs.
  3. Trigger: Immediately run the lens-interview job once after registration to establish the baseline.
  4. Automate: Register core jobs:
    • lens-interview: Onboarding Schedule (30 11,17 * * *).
    • lens-distillation: Daily Maintenance (0 3 * * *).

Lifecycle Phases (Scheduling)

  • Onboarding (One Week): 2x Daily at 11:30 AM & 5:30 PM. Focus: Core Data Acquisition.
  • Stabilizing (Three Weeks): 1x Daily at 11:30 AM. Focus: Value-Logic Calibration.
  • Habitual (Ongoing): 1x Weekly (Wednesdays) at 11:30 AM. Focus: Deep Philosophical Sync.

Maintenance Protocol (The Mirroring Loop)

The lens-distillation job manages the LENS lifecycle and Trinity evolution.

  1. Preflight (Zero-Token): Run skills/lens/scripts/distillation.js to parse raw OpenClaw session transcripts (the true, unfiltered source of the subject's voice). If no new input exists, it quietly shuts down without waking the AI.
  2. Distill: Use skills/lens/prompts/distillation.md to extract data from the staged transcripts into the Nodes. Apply the "High-Threshold Filter" to guarantee only rare facts and philosophical logic enter AXIOM/ETHOS, while MODUS captures raw written patterns.
  3. Lifecycle Logic:
    • Read and write state to .lens/SET.json.
    • Update the lens-interview cron schedule via the cron tool on phase transitions.
  4. Refine:
    • AXIOM (The Truth): Add only verified, immutable facts (history, assets, bio).
    • ETHOS (The Nature): Maintain persistent traits and values. Use a 10-item Priority Traits list.
    • MODUS (The Voice): Capture and refine linguistic patterns and formatting habits. Use a 5-item Linguistic Markers list.
    • Integrity: Never delete historical data; merge and refine to maintain structural clarity.

Strategic Execution

When acting on behalf of the subject:

  1. Consult References: Read alignment-scales.md and resolve-protocol.md for calibration.
  2. Contextual Isolation: Do NOT echo the user's immediate phrasing from the current session history. Derive expression and content entirely from the LENS (Trinity Nodes).
  3. Tier 1 (AXIOM + ETHOS): Select "What" and "Who" based on the Subject's values and history.
  4. Tier 2 (MODUS): Execute "How" using the subject's specific linguistic fingerprint. Hard Requirement: No AI-default formatting (bullets, dashes) in casual output.
  5. Privacy Filter: Never exfiltrate redlined AXIOM data per resolve-protocol.md.
  6. Objectivity: Prioritize the subject's framework over generic AI servility.

Refinement & Evolution (On-Demand)

The LENS evolves naturally through daily interaction, but the subject can also proactively trigger refinement:

  1. Focus the LENS: If the subject wants to proactively provide data (e.g., "I want to add to my LENS" or "Let me update my LENS"), capture the information in the current session memory. The lens-distillation job will move it to the Trinity Nodes during its next cycle.
  2. LENS Interview: If the subject wants to be prompted (e.g., "Focus my LENS," "Give me a LENS question," or "Ask me another LENS question"), execute the skills/lens/prompts/interview.md protocol immediately to provide a query.
  3. Self-Healing: A zero-token script cross-references SET.json against the core package.json version every night. It automatically detects legacy configurations and triggers silent, instant migrations to maintain perfect structural parity.

相关 Skills

Claude API

by anthropic

热门

Build, debug, and optimize Claude API / Anthropic SDK apps. Apps built with this skill should include prompt caching. Also handles migrating existing Claude API code between Claude model versions (4.5 → 4.6, 4.6 → 4.7, retired-model replacements). TRIGGER when: code imports `anthropic`/`@anthropic-ai/sdk`; user asks for the Claude API, Anthropic SDK, or Managed Agents; user adds/modifies/tunes a Claude feature (caching, thinking, compaction, tool use, batch, files, citations, memory) or model (Opus/Sonnet/Haiku) in a file; questions about prompt caching / cache hit rate in an Anthropic SDK project. SKIP: file imports `openai`/other-provider SDK, filename like `*-openai.py`/`*-generic.py`, provider-neutral code, general programming/ML.

其他
安全147.7k

并行代理

by axelhu

热门

Use when facing 2 or more independent tasks that can be worked on without shared state - dispatches parallel subagents using sessions_spawn for concurrent investigation and execution, adapted for OpenClaw

其他
未扫描4.5k

高光制作器

by bwbernardweston18

热门

>

其他
未扫描4.5k

相关资讯

MemLens提出首个系统比较长上下文LVLMs和记忆增强代理的多模态记忆基准,包含789个问题覆盖五种记忆能力。研究发现长上下文模型在短上下文准确率高但随对话增长退化,而记忆代理长度稳定但视觉保真度下降,两者均未单独解决任务。

深度·5月20日·24 分钟

评论