Saju Interpreter

Interpret East Asian Four Pillars / Saju charts from already-derived pillars (year, month, day, hour stems/branches). Use when the user asks for 사주, 사주팔자, 명리...

Security Scan

The skill's files, instructions, and optional pillar-calculation script align with its stated purpose (Saju interpretation); it requests no credentials and contains no obvious exfiltration or unrelated capabilities.

Purpose & Capability

Name and description (Saju interpreter) match the included SKILL.md guidance and detailed rules file. The included calculate-pillars.py is directly relevant for the explicitly allowed task of deriving pillars from a provided birth date/time; references/rules.md documents interpretation rules used by the skill. No unrelated binaries, credentials, or network endpoints are requested.

Instruction Scope

SKILL.md confines the skill to interpretation of already-derived pillars and only uses scripts/calculate-pillars.py when the user explicitly supplies calendar input and asks for derivation. The instructions do not direct the agent to read arbitrary system files, request secrets, or send data to external endpoints. The reference rules file is local and used for interpretation. The script also only outputs JSON and does not perform network I/O.

Install Mechanism

There is no install spec (instruction-only), which is low-risk. The included Python script imports korean_lunar_calendar, an external Python package; the runtime environment may need that dependency installed (pip). The absence of an install step means the runtime may fail if that package is missing—this is an operational note rather than a security issue. Verify the origin of any required Python package before installing.

Credentials

The skill declares no required environment variables, no credentials, and no config paths. The code and instructions do not access environment secrets. All required inputs (date/time or pillars) are user-provided and appropriate to the skill's purpose.

Persistence & Privilege

always is false and the skill does not request persistent system presence or modify other skills or system-wide configuration. Autonomous invocation is allowed by default but not combined with any broad privileges or secrets here.

Assessment

This skill appears coherent and focused on saju interpretation. Before installing: (1) note that the included Python script depends on the third-party package 'korean_lunar_calendar'—install it only from a trusted source if you need calendar derivation; (2) the skill does not request credentials or access files, and it prints results locally (no network calls in the provided code), but you should still review any third-party package you install for the script; (3) the interpretation rules are opinionated defaults (not an authoritative school), so consider them guidance rather than definitive judgment; (4) if you expect strict calendar conversion for edge cases (leap lunar months, zi-hour rules), manually verify the derived pillars because cultural/school-specific rules can vary.

Interpret 사주팔자를 evidence-first. Start from structure, then explain meaning. Do not jump straight to fortune-teller prose.

This skill is intentionally calm and non-gimmicky:

• prefer structure over theatrical fortune-teller language
• prefer conditional claims over deterministic prophecy
• prefer explanation over intimidation
• treat people as more complex than one neat label

Quick workflow

1. Normalize the input into four pillars.
   • Year / Month / Day / Hour
   • Each pillar = stem + branch
2. Treat pillar derivation and pillar interpretation as separate tasks.
   • If the user already gives pillars, do not re-derive them.
   • If the user gives birth date/time, use scripts/calculate-pillars.py for a first-pass calculation.
   • Explicitly mention ambiguity when relevant, especially 자시 처리, leap lunar month input, or school-specific month-boundary rules.
3. Analyze in this order:
   • 월령 / 계절성
   • 일간 / 일지 축
   • 오행 분포 + 음양 밸런스
   • 십성 분포
   • 합·충·형·해·파 / 삼합·육합
   • 조후(한난조습)
   • 신살 only as optional low-weight tags
4. Separate:
   • relationship detected
   • relationship strength
   • transformation success/failure
5. Write the interpretation in layers:
   • structural findings
   • conditional meaning
   • practical reading
   • uncertainty / school-dependent branches

Output style

Use calm, conditional language.

• Prefer: ~로 읽힌다, ~경향이 있다, ~일 때 강화된다, ~로 보는 편이 안전하다
• Avoid: 반드시, 틀림없이, 운명적으로, 무조건
• Do not present health, death, disaster, infertility, crime, or relationship doom as deterministic facts.
• Do not let 신살 overrule structure.

Default interpretation policy

Use these defaults unless the user asks for a different school.

• Priority 1: 월령(월지)와 계절성
• Priority 2: 일간 기준 십성/신강약의 기본 축
• Priority 3: 강한 삼합/방합 구조와 월지 가중
• Priority 4: 합 발견 후 합화 성립 여부 별도 판정
• Priority 5: 충은 강한 변동 요인으로 반영
• Priority 6: 형·해·파는 보정층으로 반영
• Priority 7: 신살은 보조 태그만 허용

Reading checklist

Before writing the final reading, verify:

• Did I identify the day master correctly?
• Did I anchor the reading in month branch / season?
• Did I distinguish same-element support from actual usable balance?
• Did I separate 합 from 합화?
• Did I mark any conflicting signals instead of forcing one neat conclusion?
• Did I keep 신살 below structure?

What to read when needed

Read references/rules.md when you need the detailed rule specification, including:

• 십성 decision rules
• 조후 defaults
• relation priority model
• interpretation pseudocode
• example output shape

Response template

Use a compact structure like this unless the user asks for a freer reading:

## 기본 구조
- 일간:
- 월령/계절:
- 눈에 띄는 오행:

## 관계와 포인트
- 합/충/형/해/파:
- 십성 분포 핵심:
- 조후 포인트:

## 해석
- 강점:
- 부담/리스크:
- 관계/일 방식:
- 실용 코멘트:

## 단서
- 학파나 가중치에 따라 달라질 수 있는 부분:

Boundaries

• Do not pretend there is one universally correct school.
• Do not silently invent missing pillars.
• If a pillar or birth-time assumption is uncertain, say so explicitly.
• If the user wants full derivation from solar/lunar birth info, recommend using a reliable 만세력 source first, then interpret the derived pillars with this skill. ith this skill.