LucentMind/argus-core

suspect-commits

Use when a triaged defect looks like a regression and code paths are implicated — localizes the introducing change by walking git history over the implicated paths from a stated anchor commit, ranking candidates, and recording them with author and PR in the…

原文语言：英语

contribute-back

Use when an investigation produced a reusable lesson worth adding to Argus's skills or references — a repeatable analysis procedure, a glossary/runbook correction, or a proven command recipe. Drafts an inert proposal the user reviews on the Skills page; never…

原文语言：英语

authoring-argus-packs

Use when creating, editing, or debugging an Argus pack — writing or changing an argus-pack.json manifest, or declaring a pack's persona, binaries/CLIs, evidence detectors, webPanel windows and their agent commands, skills, or references for Argus Core.

原文语言：英语

systematic-triage

Use when triaging any defect case toward a root cause — structures the investigation into evidence intake, pattern comparison, hypothesis testing, and a defensible RCA conclusion. Prevents symptom-level verdicts and evidence-free guessing. Adapted for…

原文语言：英语

code-review

Use when reviewing a linked pull request's diff — structures the review into an aggressive find pass across four lenses and a strict claim-verification pass before any finding is recorded. Prevents nitpick floods and unverified findings; a clean "no issues"…

原文语言：英语

code-graph

Use when the user asks a blast-radius, impact, "what calls X", "what breaks if", or dependency-chain question about a linked repo. Queries the pre-built graphify code graph instead of multi-hop grep exploration. NOT for general "explain this code" questions —…

原文语言：英语

demo

A demo skill.

原文语言：无法判定