Verify code drafts against requirements with up to three retry attempts.
npx skills add https://github.com/lshtram/core_dev --skill self-correction
Please help me install this Agent Skill. Skill: self-correction Source: https://github.com/lshtram/core_dev/tree/main/.agent/skills/self-correction Command: npx skills add https://github.com/lshtram/core_dev --skill self-correction
This skill helps ensure code or logic is thoroughly reviewed before it reaches the user, reducing bugs and unsafe behavior.
Provide an initial draft and a critique, then run the verification script to confirm correctness.
High-intent search queries and answers about installing and using this skill.
To autonomously critique code, provide an initial draft and critique, then run a verification tool to execute deterministic checks and return a structured self-correction report. This verifies security, edge cases, and performance before code reaches end users.
Automated code self-correction is the process of identifying and preventing flaws in logic pre-release. It works by running a verification script on draft code, returning a structured report, and supporting iterative refinements with a maximum of three retry attempts.
Yes, automated verification is applicable across AI-assisted development. You provide an initial draft and critique, then the verifier analyzes it against explicit requirements to prevent unsafe behavior and bugs before release.
Iterative code refinement supports a maximum of three retry attempts. During each attempt, the verifier runs deterministic checks on your draft and provides a structured self-correction report for the next iteration.
You need an initial code draft, a critique, and a verification tool to run checks. The verifier executes these deterministic checks on the draft and returns a structured report to guide iterative refinements.
Yes, autonomous code verification critiques security, edge cases, and performance. It analyzes your draft against explicit requirements and returns a structured self-correction report to prevent flaws before the code reaches end users.
StepowskiEricVerify post-generation outputs using external tools for factual and code validation.
ClyptAIMerge PRs, wait for CI, deploy to production, and validate canary health.
maltemdManage documentation with version control, CI/CD, and automated testing workflows.
macaughDetects injection, memory safety and authorization flaws in multiple languages.
