Track and audit belief changes with AI influence detection.
npx skills add https://github.com/TylerGarlick/abraxas --skill mnemon
Please help me install this Agent Skill. Skill: mnemon Source: https://github.com/TylerGarlick/abraxas/tree/main/skills/mnemon Command: npx skills add https://github.com/TylerGarlick/abraxas --skill mnemon
This Skill addresses the challenge of invisible AI-assisted belief revision by providing a systematic way to track your beliefs over time, record their evolution, and identify potential AI influence.
Use the mnemon skill to hold the belief that "the sky is blue" with high confidence.
High-intent search queries and answers about installing and using this skill.
Belief tracking to detect AI influence works by logging your initial stance with confidence levels, recording subsequent revisions with attribution, and correlating those revision events with AI interaction timestamps to flag potential sycophancy patterns.
AI sycophancy detection identifies beliefs that shifted shortly after AI interaction. Belief auditing records the type of revision—such as strengthened, weakened, or replaced—along with the reason, providing a detailed historical analysis of your evolving views.
You log and audit belief revisions over time by registering your initial belief states with supporting evidence, then recording each change event with its specific revision type and attribution to analyze whether updates were driven by new evidence or AI suggestions.
Belief tracking is suited for complex research topics where your views evolve. You can log your initial stance, track how it changes after reading articles or interacting with an AI, and audit whether those changes were driven by new evidence or undue AI influence.
Structured storage for belief states and revision events is required to correlate historical changes with AI interaction timestamps. This environment setup allows the system to manage belief registration, revision logging, and detailed attribution analysis effectively.
Auditing whether belief changes were driven by evidence or AI suggestions addresses the challenge of invisible AI-assisted belief revision. This process systematically tracks your beliefs over time to identify potential sycophancy and undue algorithmic influence on your decisions.
GuitarAlchemistManage belief states using tetravalent logic with confidence levels.
GuitarAlchemistFuse AI annotations and research claims into temporal assumption graphs.
Soul-Brews-StudioParse user affirmations and log belief records with timestamps and context.
Mathews-TomStructure multi-step reasoning with numbered steps, revision checkpoints, and branch exploration.
