A skill engineered for when you need your agent to see the world through your LENS. It evolves through three core nodes to ensure every interaction reflects your unique identity.
Prompt (The Request) + LENS (The Trinity Nodes) = Authentic Output.
A digital shadow should not be a burden to maintain. Most persona systems require manual adjustment and complex prompt-engineering. LENS is a passive, effortless implementation that grows sharper with use.
By observing interactions and distilling the essence of identity, the system mirrors the subject with zero-point objectivity. The architecture is token-efficient, separating identity into three surgical nodes to ensure the agent only carries the context required for authentic manifestation.
Complexity is mapped into three distinct nodes. These anchors define the subject's identity through specific, immutable parameters.
History and reality form the Cognitive Bedrock. Identity is grounded in the immutable facts of existence.
Chronology, Kinship, Assets, Professional Credentials, and Immutable History.
Values and character form the Internal Compass. Identity is defined by how the world is weighed.
Decision-Logic, Aesthetic Triggers, Philosophical Alignment, and Priority Traits.
Style and expression form the Behavioral Interface. Identity is rendered through the nuance of delivery.
Cadence Velocity, Punctuation Density, Lexical Gravity, Tone Temperature, and Linguistic Markers.
To keep the Trinity Nodes pure, the LENS uses an isolated configuration layer:
SET.json: Manages the "State Machine." This file tracks your lifecycle phase, question counts, and model preferences. This separation ensures that skill updates never overwrite your identity or your settings.
The LENS is not static; it is an evolving digital shadow.
- The Interview Protocol: The system progresses through onboarding (2x daily for 1 week), stabilizing (1x daily for 3 weeks), and habitual (1x weekly) schedules to calibrate your digital shadow.
- The Distillation Loop: The engine of evolution. A background routine analyzes patterns to maintain structural peace without user intervention.
- Contextual Isolation: To prevent "AI-muddiness," the distillation engine strictly analyzes the subject's raw input, ignoring its own previous outputs.
- Silence Toggle: You have total control. Setting
questionstofalseinSET.jsonsilences the interview routine while keeping the identity engine active. - Refinement on Demand: You can proactively sharpen the LENS at any time. Simply say "Focus my LENS" or "Ask me a LENS question" to trigger an immediate calibration.
- Self-Healing Architecture: Starting with v0.6.3, the system is equipped with an automated repair engine. It detects legacy configurations or missing environment variables and triggers silent migrations to ensure your LENS remains performant and up-to-date without manual intervention.
LENS utilizes three flat Markdown files and a single JSON state file instead of a complex vector store. This approach is lightweight, token-efficient, and human-readable. By treating identity as Operating System nodes, we have made the perspective of the subject computable without the need for heavy databases.
Proudly refined through my own LENS.