Classify requests into repo-grounded, doc-grounded, or external-facts and enforce verification.
npx skills add https://github.com/drsapaev/final --skill aif-grounded-drsapaev
Please help me install this Agent Skill. Skill: aif-grounded Source: https://github.com/drsapaev/final/tree/main/.codex/skills/aif-grounded Command: npx skills add https://github.com/drsapaev/final --skill aif-grounded-drsapaev
This Skill acts as a critical reliability gate, preventing AI from generating speculative or fabricated information by enforcing strict evidence-based reasoning.
Use the aif-grounded skill to answer the question about the latest policy document, ensuring the answer is fully verified.
High-intent search queries and answers about installing and using this skill.
To prevent AI hallucinations and ensure responses are evidence-based, this skill enforces strict verification by classifying requests into repo-grounded, doc-grounded, or external-facts, and mandates a confidence gate before providing answers.
Uncertainty handling for high-stakes information works by explicitly stating when data is missing or confidence is below 100%, ensuring the AI avoids guesswork and clearly flags knowledge gaps instead of fabricating speculative answers.
The best way to verify changeable facts before implementing code changes is to apply a confidence gate that mandates verification against provided context, ensuring all actions are fully supported by evidence-based reasoning.
Yes, you can use this skill to validate answers against repository documentation by classifying requests as doc-grounded, enforcing that all generated responses are directly supported by the provided context.
You need explicit grounding for AI-generated answers when dealing with high-stakes information like medical guidelines, where the skill acts as a reliability gate to prevent fabricated information and guarantee verified facts.
The AI explicitly states when information is missing or confidence is low to implement explicit uncertainty handling, preventing guesswork and ensuring that users are aware of any knowledge gaps before relying on the provided answer.
LaiTszKinCombine repository discovery with web research to answer questions with sourced evidence.
vulikjulikRequire verifiable evidence before answering fact-sensitive requests.
nimadorostkarOptimize retrieval-augmented generation pipelines with hybrid search, reranking, and semantic chunking.
ntaksh42Detect and correct errors in AI outputs with automated validation reports.
