Convert debugging outcomes into reusable codebase-specific skills with quality validation.
npx skills add https://github.com/marcmunoz-uno/jailbreak --skill learner-marcmunoz-uno
Please help me install this Agent Skill. Skill: learner Source: https://github.com/marcmunoz-uno/jailbreak/tree/main/skills/learner Command: npx skills add https://github.com/marcmunoz-uno/jailbreak --skill learner-marcmunoz-uno
It helps turn a complicated, codebase-specific win (a tricky bug fix, a hidden gotcha, or an undocumented behavior) into a reusable skill that can be applied to future problems instead of being forgotten.
Ask the AI to extract a learner skill by providing the specific error message, the exact fix (with file paths/line numbers), and the recognition triggers from your latest debugging session.
High-intent search queries and answers about installing and using this skill.
You extract reusable skills from debugging outcomes by providing the specific error message, exact fix with file paths, and recognition triggers to generate a structured, codebase-relevant skill for future problem solving.
Capturing project learning for undocumented behaviors requires classifying the outcome into either updateable expertise or stable workflow, applying a quality gate to ensure the knowledge is precise and actionable rather than vague.
A quality gate improves knowledge capture by rejecting overly generic, easily Googleable, or vague learnings, ensuring that only precise, codebase-specific gotchas and fixes are saved as reusable skills.
You should classify a debugging fix as expertise instead of a workflow when the outcome involves updateable knowledge, whereas stable, repeatable processes should be classified as workflow to improve the safety of future improvements.
To start extracting a skill from a codebase-specific gotcha, you need the specific error message encountered, the exact fix including file paths and line numbers, and the recognition triggers that identify the problem.
No, this skill extraction framework applies a quality gate that rejects generic, easily Googleable errors, focusing solely on non-obvious gotchas, undocumented behaviors, and fixes that required real investigation.
ZhiruiFengAnalyze codebases to map modules, dependencies, and architectural patterns.
d-o-hubExtract non-obvious session learnings and persist them into MEMORY.md.
OmerZeyveliRecords validated UI R&D findings from complex debugging sessions into searchable, structured summaries.
jimmi2051Capture conversation insights into reusable skills with YAML frontmatter and Workflow format.
