- Home
- Skills
- Leegonzales
- Aiskills
- Silicon Doppelganger
silicon-doppelganger_skill
- Python
20
GitHub Stars
4
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill leegonzales/aiskills --skill silicon-doppelganger- CHANGELOG.md2.0 KB
- LICENSE1.0 KB
- README.md8.4 KB
- SKILL.md8.7 KB
Overview
This skill builds psychometrically accurate personal proxy agents ("Digital Twins") for the PAIRL Conductor system. It extracts personality, decision heuristics, and values into a portable persona schema that lets agents negotiate, filter, and act on a principal's behalf. The output is a Conductor-ready persona profile that preserves fidelity across sessions and systems.
How this skill works
Run a structured workflow: a guided interview collects psychometrics, linguistic samples, narrative identity, and decision heuristics. The data is encoded into a compact XML persona schema that defines core drivers, decision logic, agent rules, and stress behaviors. The proxy is validated via behavioral testing against the principal and then integrated as a spoke in PAIRL Conductor with explicit must_reject, must_protect, and should_prefer rules.
When to use it
- Create a personal proxy agent to accept or reject tasks automatically
- Build a Digital Twin for PAIRL Conductor integration and calendar negotiation
- Extract robust personality and decision patterns for agent representation
- Validate an AI proxy against real-world principal choices
- Model team dynamics by simulating multiple proxies
Best practices
- Conduct a 45–60 minute structured interview and collect concrete examples and quotes
- Capture psychometrics (CliftonStrengths, VIA) and communication samples for a linguistic fingerprint
- Encode explicit agent rules (must_reject, must_protect, should_prefer) with clear thresholds
- Run a question battery and target ≥80% lenient-match accuracy before deployment
- Document contradictions, shadow behavior, and blind spots rather than smoothing them
Example use cases
- Proxy negotiates calendar slots and declines requests that violate energy or craft boundaries
- Team leads simulate partnership friction by loading multiple persona schemas
- Individuals surface blind spots and produce a ‘how I work’ artifact for collaborators
- WritingPartner uses the linguistic fingerprint to calibrate tone and phrasing
- Deploy a proxy to filter incoming email and surface only aligned requests
FAQ
Aim for 80%+ lenient-match accuracy on a behavioral question battery; refine mismatches iteratively.
What artifacts are produced?
A Conductor-ready XML persona schema, an origin-story document, extraction checkpoints, and validation question sets.