Audit generated code against implementation specs and security standards.
npx skills add https://github.com/Pyl-Tech/stream-coding --skill code-review-pyl-tech
Please help me install this Agent Skill. Skill: code-review Source: https://github.com/Pyl-Tech/stream-coding/tree/main/.agents/skills/code-review Command: npx skills add https://github.com/Pyl-Tech/stream-coding --skill code-review-pyl-tech
Prevents shipping incorrect behavior by verifying generated code conforms to the written spec, security rules, and language standards, with a Golden Rule that the spec is corrected before code.
Ask an AI agent to perform a conformance audit of your pull request using the Stream Coding code-review skill, focusing on Golden Rule divergence and producing the structured report with verdict.
High-intent search queries and answers about installing and using this skill.
A spec-to-code conformance audit verifies that generated code matches the implementation spec, security rules, and language standards by mapping each requirement to exact YES/NO/PARTIAL implementation status, treating divergence as a spec issue first.
Perform a semantic code review by establishing review scope from PR metadata, running automated diagnostics like typecheck and lint, reading full file contents, mapping requirements to implementation status, and generating a divergence-first structured report with a final verdict.
Semantic verification is needed for PR and local diff review workflows when structural checks are insufficient and you must verify actual behavior across the full module scope to ensure generated code conforms to the written spec.
Yes, the code review process checks universal CRITICAL and HIGH security items including auth bypass, injection, and XSS, while applying a language-specific review checklist to ensure required security standards are met.
Verify test quality and CI readiness by establishing review scope from git diff context, confirming merge readiness using PR metadata, applying evaluator separation safeguards, and mapping requirement implementation status including test wiring.
When generated code diverges from the implementation spec, the Golden Rule dictates that the spec is corrected before the code, and the divergence is reported first as a spec issue through a structured divergence-first reporting workflow.
EmasoftRetrieve PR metadata, files, and merge status as structured JSON.
arbazkhan971Assess code or feature specifications with multi-persona risk verdicts and voting gates.
uesleinaschIdentify and report defects, security, and architectural concerns in code.
Ilenburg1993Audit code for semantic flaws and generate a Markdown report with corrections.
