Domains
Forms
Mechanism
Breaks down how the skill works and what it produces, so you can quickly judge whether it fits your scenario and value.
When starting a fine-tuning effort, converting traces into an eval set, or calibrating a judge against human labels. Required before any training config is written.
An eval/ directory with goldens, graders, drift suite, and baseline token that gates every downstream fine-tune run and checkpoint promotion.
Use Cases
A team plans to fine-tune a support model but has only production chat logs. Using this skill they open-code 100+ traces into failure buckets, write deterministic graders, calibrate an LLM-judge for tone, and produce baseline-<model>.json. This gives them labeled training data and a regression gate, preventing silent quality drops after training.
A researcher has no real data and synthesizes goldens by dimension sampling. They follow the skill to build graders, freeze a drift suite, and establish a base-model baseline before method selection. Later checkpoint-promotion re-runs the same harness, catching regressions objectively against the frozen baseline.
Skill Relationships
Dependency relationships read as "the upper tier points to the lower tier." The current Skill sits in the middle tier — above are Skills that depend on it, below are Skills it depends on.
Tier 1 · These Skills Use Me
Tier 2 · Current Skill
Tier 3 · I Use These Skills
Browse this skill's relationships within its skillset. Click a node to switch the side panel; use the search box to jump to any skill.
Skill File
