Compare agent prompt definitions against test harnesses for structural consistency.
npx skills add https://github.com/whichguy/skill-craft --skill prompt-align
Please help me install this Agent Skill. Skill: prompt-align Source: https://github.com/whichguy/skill-craft/tree/main/plugins/prompt-align/skills/prompt-align Command: npx skills add https://github.com/whichguy/skill-craft --skill prompt-align
This skill resolves discrepancies between agent or skill prompts and their corresponding test harnesses, ensuring that phase models, skip conditions, and wiring remain synchronized during development.
Run the prompt-align skill by providing the path to your agent or skill directory to automatically detect and compare it against its associated test harness.
High-intent search queries and answers about installing and using this skill.
To ensure prompt consistency across agent test harnesses, you can compare agent prompt definitions against their respective test harnesses to verify structural and logical alignment in phase tables, wiring, and skip conditions by providing the repository root path.
Silent test failures occur when agent phase logic updates are not synchronized with the test harness. Comparing phase models, skip conditions, and wiring lists identifies mismatches and provides actionable fix recommendations to resolve these discrepancies.
The best way to audit agent prompts against test harnesses is to perform automated diffing of phase tables and wiring lists, which identifies authoritative sources for conflicting logic and generates actionable remediation reports.
To detect mismatches in phase tables and skip conditions, run an automated harness discovery from your repository root to compare structural definitions against the test suite and generate a remediation report.
Yes, you need repository root access to perform automated harness discovery and accurately compare agent or skill prompt definitions against their corresponding test harness structures.
hiddink-aiValidate agent-to-skill and agent-to-guide symlink mappings in hiddink-harness.
vm0-aiScan codebases for quality issues and generate GitHub tracking issues.
nounderTest Effect.ts success and failure paths with deterministic concurrency handling.
MuzeWinterGather definitions, decisions, and source artifacts for analytical questions.
