Write root cause blueprint
Skill YuanshengClaw/yuansheng-kit/plugins/trace/skills/write-root-cause-blueprint
Agent-oriented toolkit for RISC-V optimization knowledge extraction, pattern mining, real-hardware performance analysis, root-cause diagnosis, and code generation
npx -y skills add YuanshengClaw/yuansheng-kit --skill write-root-cause-blueprintAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
3 things to look at
- 16 days oldThe repository was created 16 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Produce and semantically review evidence-grounded Yuansheng Root Cause Blueprint v1-lite candidates from validated perf evidence and a confirmed hardware profile. Use when Yuansheng Trace requests a per-function Blueprint candidate or an independent review of one.
SKILL.md
1.5 KB, as published. Nobody here has run it
Write a Root Cause Blueprint
- Bind the candidate to the supplied software, test case, ranked function, and confirmed hardware profile.
- Use only the supplied perf statistics, annotate output, metadata, hardware profile, and their recorded hashes. Do not read source code or acquire new evidence.
- Distinguish observations from hypotheses. Copy measured values exactly and
keep unsupported facts as
nullwith the required machine-readable gap codes. - Keep every
rvv_patternandpattern_confidencevaluenull, and includepattern_catalog_unavailable. Do not name or infer a Pattern. - Do not invent a source file, line, code location, metric, baseline, or hardware capability. Keep recommendations within what the evidence can support.
- Set
needs_human_reviewtotrueandallow_auto_forward_to_ys_crafttofalse. - Return strict JSON conforming to the supplied v1-lite Schema and preserve the workflow-requested claim-to-evidence references.
For semantic review, compare every material claim with the immutable candidate and supplied evidence. Report unsupported, conflicting, or uncited claims without modifying the candidate. Require a new candidate and full validation when correction is necessary.