How does Epharmoge differ from standard AI code generation?

Standard generation focuses on correctness. Epharmoge adds a verification layer to check if that 'correct' code fits your local conventions, dependencies, and environment.

Can Epharmoge fix the mismatches it finds?

Yes, once a mismatch is identified and confirmed by the user, the protocol enters an adaptation phase to modify the result until it fits the context.

When should I use the /contextualize command?

Use it after the AI has performed a task if you want to verify that the output respects your project's specific constraints or if the AI's assumptions seem misaligned with your local setup.

What is the Epharmoge skill in Claude Code?

Epharmoge is a protocol that ensures AI-generated results are applicable to your specific project context, bridging the gap between theoretical correctness and practical utility.

What dimensions of context does this skill check?

It evaluates mismatches across four primary dimensions: Convention (coding standards), Environment (infrastructure/state), Audience (user needs), and Dependency (external requirements).

Contextual Applicability (Epharmoge)

Name: Contextual Applicability (Epharmoge)
Author: jongwony

byjongwony

•

生産性とワークフロー

Verifies that technically correct AI outputs fit the specific application context, environment constraints, and local project conventions.

Epharmoge is a specialized epistemic protocol designed to detect mismatches between an AI's execution and the actual application context. While an AI might generate technically 'correct' code or analysis, it often misses subtle local requirements like specific project conventions, environment constraints, or architectural patterns. This skill surfaces these mismatches for user judgment, ensuring that every result is not just accurate in a vacuum, but warranted and applicable to your specific codebase and organizational goals. It prioritizes applicability over mere correctness through a structured dialogical process.

主な機能

0188 GitHub stars

02Verifies 'warranted assertibility' to ensure outputs fit the environment's unique constraints.

03Provides structured mismatch tracking through automated task creation and status updates.

04Detects mismatches between technical output and local project conventions or environment states.

05Supports iterative adaptation of results based on user-directed context and feedback.

06Surfaces specific evidence for applicability gaps before finalizing execution.

ユースケース

01Ensuring infrastructure-as-code scripts match specific environment states or cloud provider constraints.

02Checking if documentation or analysis is correctly scoped for the intended audience or specific project phase.

03Validating that generated code follows local styling or architectural patterns not found in global training data.

主な機能

0188 GitHub stars

02Verifies 'warranted assertibility' to ensure outputs fit the environment's unique constraints.

03Provides structured mismatch tracking through automated task creation and status updates.

04Detects mismatches between technical output and local project conventions or environment states.

05Supports iterative adaptation of results based on user-directed context and feedback.

06Surfaces specific evidence for applicability gaps before finalizing execution.

ユースケース

01Ensuring infrastructure-as-code scripts match specific environment states or cloud provider constraints.

02Checking if documentation or analysis is correctly scoped for the intended audience or specific project phase.

03Validating that generated code follows local styling or architectural patterns not found in global training data.