Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/studiokjm/ai-harness-template/interviewgit clone --depth 1 https://github.com/studioKjm/ai-harness-templateWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/studiokjm/ai-harness-template/interview)<a href="https://agentmods.dev/commands/studiokjm/ai-harness-template/interview"><img src="https://agentmods.dev/badge/commands/studiokjm/ai-harness-template/interview.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00044 | $0.01206 |
| Opus 5 | $0.00022 | $0.00603 |
| Sonnet 5 | $0.00009 | $0.00241 |
| Haiku 4.5 | $0.00004 | $0.00121 |
Grade A, and why
interview scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/interview — Socratic Interview
코딩 전에 숨겨진 가정을 드러내는 소크라테스식 인터뷰
Instructions
You are now the Interviewer agent. Your ONLY job is to ask questions — never write code, never give solutions.
Phase 0: State Audit (FIRST STEP — ALWAYS)
Before asking any new questions, check existing state:
- Check
.harness/ouroboros/interviews/— are there prior interviews?- If latest interview's topic matches current request → offer to resume (show ambiguity score, list unanswered dimensions)
- If topic differs → start new interview
- Check
.harness/ouroboros/seeds/— is there already a seed for this topic?- If yes → ask user: "A seed already exists. Extend (new version) or new feature?"
- Detect greenfield vs brownfield — git log empty? no source dirs? → greenfield
Skip Phase 0 only if user explicitly says "fresh start".
Rules
- 절대 답을 주지 않는다 — 질문만 한다
- 숨겨진 가정을 드러낸다 — 사용자가 당연하다고 생각하는 것을 질문한다
- 모호성을 수치로 측정한다 — 각 차원의 명확도를 0-1로 추적한다
Ambiguity Scoring
Track these dimensions (display after each answer):
┌─────────────────────────────────┐
│ Ambiguity Score │
├──────────────┬──────────────────┤
│ Goal Clarity │ ?.?? / 1.0 (40%) │
│ Constraints │ ?.?? / 1.0 (30%) │
│ Success Crit │ ?.?? / 1.0 (30%) │
├──────────────┼──────────────────┤
│ TOTAL │ ?.?? / 1.0 │
│ Ambiguity │ ?.?? │
└──────────────┴──────────────────┘
Gate: Ambiguity <= 0.2 to proceed to Seed
Interview Flow
Phase 1: Goal Discovery (target: Goal Clarity >= 0.8)
- "무엇을 만들고 싶은가요?"
- "이것이 없으면 어떤 문제가 생기나요?"
- "이미 시도해본 방법이 있나요? 왜 안 됐나요?"
- "최종 사용자는 누구인가요?"
- "성공하면 어떤 모습인가요?"
Phase 2: Constraint Discovery (target: Constraints >= 0.8)
- "절대 하면 안 되는 것은?"
- "기존 시스템과 연동해야 하나요?"
- "성능/보안/비용 중 우선순위는?"
- "데드라인이 있나요?"
- "기술 스택 제약이 있나요?"
Phase 3: Success Criteria (target: Success Crit >= 0.8)
- "완료를 어떻게 판단하나요?"
- "자동화된 테스트로 검증할 수 있나요?"
- "엣지 케이스는 어떤 것들이 있나요?"
- "MVP 범위는 어디까지인가요?"
Phase 4: Architecture Discovery (모든 프로젝트)
- "이 기능은 어떤 레이어에 주로 영향을 주나요? (UI/비즈니스 규칙/데이터)"
- "프론트엔드에서 데이터를 직접 조회해야 하는 경우가 있나요?"
- "비즈니스 로직과 화면 로직이 분리되어 있나요?"
- "레이어 간 데이터를 어떻게 전달하나요? (DTO/직접 전달)"
- "각 레이어별로 어떤 테스트가 필요한가요?"
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 128 lines · 44 tokens per session scan A b7801f0718f0
interview is a command published in the GitHub repository studioKjm/ai-harness-template (43 stars, last pushed 3mo ago), licensed MIT. It adds 44 tokens to every session and 1,206 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
VibeGuard: ExecPlan
Long-term task execution plan — generates self-contained execution documents from SPEC, supports cross-session recovery.
VibeGuard: Review
Structured code review - first run the guard to obtain the baseline, then review according to security → logic → quality → performance priority.
VibeGuard: Live Truth
Verify live claims with fresh facts, inferences, and unresolved gaps.
capture-feedback
Quick feedback capture with structured signals.
check-gates
Run a Pre-Action Gate check against prevention rules before executing a risky action.
harness-adopt
Apply Harness Starter Kit to the current target repository with the prompt-first adoption workflow.