
from immortal-skill608
Distills a person’s multi-platform chat and document history into a structured, evidence-graded digital twin skill across seven role templates (self, colleague,
Digital Immortality (Immortal Skill) ingests multi-platform data — chat logs, social archives, documents — and extracts structured persona artifacts across seven role templates (self, colleague, mentor, family, partner, friend, public figure). It guides agents through ethical intake, multi-source collection, dimensional extraction (procedural behavior, interaction patterns, memories, personality), evidence grading, conflict resolution, and packaging into loadable Agent Skill manifests. The focus is on traceable evidence tiers (verbatim, artifact, impression) and explicit ethical guardrails for sensitive roles.
Use when a user asks to distill X, create a digital twin, retain someone’s style/memories, or provides collected materials and requests a persona-based AI skill. It is appropriate for team onboarding (colleague/mentor), personal memory preservation (self/family), and public-figure summarization (public sources only) — with role-specific ethics checks before collection.
Compatible with general LLM agent runners that can run Python-based collection/manifest tooling and Read/Write tools (Cursor, Copilot-style agents, and custom OpenClaw integrations).
Digital persona distillation framework that guides agents through 8 phases (role selection → ethics → data collection → extraction → merge → output → stamp → notify) to create evidence-graded digital twin skills. Well-structured SKILL.md with clear phase progression and built-in ethics safeguards, but all referenced CLI tools (immortal_cli.py, manifest_tool.py, version_tool.py) and template files (personas/, recipes/, prompts/) are missing from the bundle — the skill cannot function without them. No scripts were available to run.
Interesting and well-thought-out concept with strong ethics guardrails built into the workflow. The SKILL.md is comprehensive with 8 clear phases, evidence grading, conflict resolution, and version management. However, the skill is essentially a framework without its actual tools — all the Python scripts and template files it references are missing from the bundle. This makes it impossible to use as-is. Security is good: no destructive commands, no exfiltration risks, explicit ethics phase, and clear 'don't do' list. Architecture follows skill spec reasonably well with frontmatter and phased structure, but the missing dependencies significantly impact both quality and usefulness scores.