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.
After a training run produces a checkpoint, when deciding whether a tuned model ships, or when a promoted model needs re-gating against updated goldens.
promotion-report.md with terminal PROMOTE/REJECT and one top remediation.A gated verdict with evidence: unsafe checkpoints are REJECTed before shipping; promoted models are cleared for quantized-export.
Use Cases
A team finishes LoRA fine-tuning a support model and gets a checkpoint that beats the task metric. Before shipping, they run this skill: it checks for golden leakage, diffs the drift suite against baseline, runs a paired arena vs base, and outputs promotion-report.md with REJECT due to 6pt general drift. The team avoids deploying a model that lost core capability, and follows the replay-mix fix.
A previously promoted model is re-gated against updated goldens after product scope change. This skill re-runs stage 2 and 3 with the new drift suite, finds a 3pt drift in the 2–5 band, triggers a seed-variation rerun that lands at 0.8pt, and emits PROMOTE. The re-gating gives confidence the model still meets quality under new requirements.
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