Weigh Git PRs and MRs into a four-axis verdict from diffs and metadata.
npx skills add https://github.com/LarryHsiao/Skadi --skill mithrandir
Please help me install this Agent Skill. Skill: mithrandir Source: https://github.com/LarryHsiao/Skadi/tree/main/skills/mithrandir Command: npx skills add https://github.com/LarryHsiao/Skadi --skill mithrandir
Lindir weighs PRs/MRs against a four-axis framework to provide actionable guidance to editors and counsel. The verdict structure balances cohesion, proportion, direction, and risk to help decide if a change should merge, hold, or be refused.
Run /mithrandir <url> to weigh a PR or MR and render a four-axis verdict.
High-intent search queries and answers about installing and using this skill.
Weighing a pull request for a merge decision involves evaluating diffs and metadata across a four-axis model covering cohesion, proportion, direction, and risk to render a structured verdict on whether to merge, hold, or refuse the change.
Automating code review impact analysis is best handled by weighing PRs or MRs against a four-axis framework, which renders a compact summary and optional action lists to guide editors on whether to merge, hold, or refuse.
Yes, you can evaluate Git merge requests directly from a URL by supplying the link to weigh the PR or MR, rendering a concise four-axis verdict from the fetched diffs and metadata.
Yes, PR weighing works with branch-local diffs by evaluating the changes against the four-axis framework to provide a consistent assessment of change impact for code review workflows.
Limitations of using a four-axis model for code review include its focus on cohesion, proportion, direction, and risk, which provides a structured verdict but may not capture nuanced context outside the diff and metadata.
maxmurrGenerate markdown MR/PR descriptions from git branch changes.
ghostwrightMonitor user promises from conversations and check follow-through.
jMertaConvert ambiguous bug reports into reproducible repros, root-cause explanations, and minimal fixes.
jh941213Spawn an independent Evaluator agent to score code artifacts across four axes.
