cheat-predict

Generate immutable blind-prediction logs with YAML headers and hook-based locking.

12|Updated May 29, 2026

Search Tags:#logging #calibration #phases #immutable #task-review #ai-analysis #blind-prediction

Ready to run

npx skills add https://github.com/Jason5330/ai-self-eval --skill cheat-predict

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: cheat-predict
Source: https://github.com/Jason5330/ai-self-eval/tree/main/skills/cheat-predict
Command: npx skills add https://github.com/Jason5330/ai-self-eval --skill cheat-predict

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Writing task-specific predictions that are immutable once created, enabling reliable retrospectives and preventing post-hoc changes during calibration loops.

Core Features & Use Cases

Immutable blind-prediction logs that are created from a given script and rubric, then locked by a hook.
Phase-driven workflow (Phase 0.5 through Phase 7) with optional review and arbitration to ensure correctness.
Integrates user reviews to produce final predictions and associated artifacts for archival.

Quick Start

Provide a target script path and rubric notes to initiate an immutable blind-prediction cycle; the system will generate a draft for review and then finalize upon approval.

Frequently Asked Questions about cheat-predict

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I create immutable blind predictions for task calibration?▼

Immutable blind predictions are generated from a target script path and rubric notes, creating a draft log that is locked via a hook after submission to prevent post-hoc changes during calibration loops.

What is a blind-scoring sub-agent and how does it work with task reviews?▼

A blind-scoring sub-agent evaluates the prediction draft objectively and integrates with a user review to finalize and record the prediction, ensuring correctness through arbitration across phase-driven workflow steps.

How do I enforce frontmatter constraints and auditability in prediction logs?▼

Auditability is enforced by applying strict frontmatter constraints, including a YAML header with script path, script hash, calibration state, and a comprehensive BlindScore header for reliable retrospectives.

Can I use blind prediction logging for coding sessions and learning topics?▼

Yes, immutable blind-prediction logging applies to coding sessions, learning topics, and mission briefs, generating task-specific predictions through a phase-driven workflow from Phase 0.5 through Phase 7.

Why do my task predictions need to be locked after submission?▼

Predictions need to be locked after submission to prevent post-hoc changes, enabling reliable retrospectives and maintaining strict auditability during calibration loops.

What is the best way to finalize and record blind predictions for archival?▼

The best way to finalize and record blind predictions is through a phase-driven workflow that integrates user reviews to produce final predictions and associated artifacts for archival.