Phdtaketaketake

Evidence-first PhD advisor matcher — scores connection strength on a 4.0 scale, not h-index.

Autor: powerofjinbo · powerofjinbo/phdtaketaketake

Testowano · Działa ★ 9.6/10

Phdtaketaketake — Evidence-first PhD advisor matcher — scores connection strength on a 4.0 scale, not h-index.

Co robi ten skill

Ranks candidate PhD advisors for a student profile using a deterministic 5-pillar (Connection/Advisor/Publication/Experience/GPA) scoring pipeline, with a hard 'no fabricated evidence' contract enforced by a --strict-evidence flag. Triggers on requests to score PhD application chances, find matching advisors, or rank professors by connection-first fit; also ships a parallel LaTeX CV-formatting sub-workflow.

Raport z testu

Ran the bundled physics_hep_audit_demo fixture through scripts/match.py for real (fresh venv, pydantic+pyyaml only) and got a deterministic, confidence-banded ranking (application_strength 3.30/2.76/2.68, bands ±0.58) with a full missing/unsourced-signal audit — a baseline unaided ranking has no such quantified uncertainty and would very likely cite unverifiable connections as fact, which is exactly what this skill's evidence contract forbids.

Testowano: 2026-07-14 · Claude Code 2.x (agent harness)

Instalacja

git clone https://github.com/powerofjinbo/phdtaketaketake
cd phdtaketaketake
mkdir -p ~/.claude/skills
cp -r . ~/.claude/skills/phdtaketaketake

Komendy i przykładowe prompty

/phdtaketaketakeEvidence-first PhD advisor matcher — scores connection strength on a 4.0 scale, not h-index.

Skille uruchamiają się na zwykłe polecenia — bez komend do zapamiętania. Po instalacji aktywują go prompty takie jak te (po angielsku):

Match me with PhD advisors who actually fit my research
Score how well this professor's work connects to my topic
Find supervisors for my HEP thesis with evidence, not rankings

Pełny raport →