Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/kimsanguine/hplannpx agentmods add commands/kimsanguine/hplan/harness-buildWrote 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/kimsanguine/hplan/harness-build)<a href="https://agentmods.dev/commands/kimsanguine/hplan/harness-build"><img src="https://agentmods.dev/badge/commands/kimsanguine/hplan/harness-build.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 5.1 | $0.00081 | $0.07523 |
| Opus 5 | $0.00041 | $0.03762 |
| Sonnet 5 | $0.00016 | $0.01505 |
| Haiku 4.5 | $0.00008 | $0.00752 |
Grade A, and why
harness-build 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 7d 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 — 607 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/harness-build
빌드 전체 파이프라인 — 게이트 → PRD → 스프린트 → 설계 → 추적
Routing
$ARGUMENTS에서 --step 플래그를 파싱한다.
| 플래그 | 실행 범위 |
|---|---|
--step evidence |
Evidence Gate만 실행 |
--step product |
Product Gate만 실행 |
--step cogs |
COGS Sentinel만 실행 |
--step prd |
PRD 작성만 실행 (checkpoint 승인 필요) |
--step okr |
OKR 설계만 실행 |
--step sprint |
Sprint 플래닝만 실행 |
--step craft-init |
디자인 시스템 초기화만 실행 |
--step craft-lint |
디자인 품질 검사만 실행 |
--step track-init |
진척 추적 초기화만 실행 |
--step track-status |
라이브 진척 현황만 실행 |
--step track-retro |
회고 및 TK 추출만 실행 |
--step scope |
범위 가드 검사만 실행 |
--step spec-review |
Phase 7(Spec Compliance Review)만 실행 |
--step quality-gate |
Phase 8(Quality Gate)만 실행 |
| 플래그 없음 | Phase 1(evidence) → 2(product) → 3(COGS) → 4(PRD+sprint) → 7(spec-review) → 8(quality-gate) 전체 실행 |
플래그 뒤의 나머지 텍스트가 대상 브리프입니다.
Instructions
You are running the hplan Build Pipeline for: $ARGUMENTS
Phase 1 — Evidence Gate (--step evidence)
왜: 아이디어를 코드로 만들기 전, 실제 고객 고통이 존재하는지 먼저 확인합니다. 이 단계를 건너뛰면 아무도 원하지 않는 것을 완벽하게 만드는 문제가 생깁니다.
Step 1-1: Exclusion Collision Check
python3 hplan/scripts/exclusions_registry.py check "$TARGET"
- COLLISION (reopen_trigger 미충족) → 즉시 STOP. COLLISION 상세 + reopen_trigger 조건 출력.
- CLEAR → 다음 단계 진행.
Step 1-2: Input 구조화
다음 항목을 수집한다: idea, icp_segment, recent_event, workaround_tool, monthly_cost_estimate, alternatives, interview_notes (줄당 1개), interview_count.
Step 1-3: 100-Point Evidence Rubric
evidence-rubric 스킬 적용 — scripts/generate_report.py 실행. 점수 + 취약 축 캡처.
Evidence Rubric: 아이디어 신뢰도를 8개 축·100점으로 채점하는 기준표. GO(75+) / INVESTIGATE(55-74) / HOLD(55 미만) 3단계로 판정합니다.
Step 1-4: Interview Synthesis
interview_notes가 부족하거나 decision == "interview"이면 interview-synthesis 스킬 적용. AI export 인제스트 또는 신규 인터뷰 계획. 5/3 strong-Push 룰 감사.
Step 1-5: Evidence Decision
| 점수 | 추가 조건 | 결정 | 의미 |
|---|---|---|---|
| ≥ 75 | interview_lines ≥ 2 AND economic pain 있음 | build |
Product Gate로 진행 |
| ≥ 75 | interview_lines < 2 OR economic pain 없음 | interview |
점수 강하지만 증거 부족 |
| 55–74 | — | interview |
인터뷰 추가 필요 |
| 35–54 | — | pivot |
문제 정의 재검토 |
| < 35 | — | hold |
STOP — 증거 불충분 |
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.
- 7d ago First seen · 607 lines · 81 tokens per session scan A 439439bf0c97
harness-build is a command published in the GitHub repository kimsanguine/hplan (2 stars, last pushed 22d ago), licensed MIT. It adds 81 tokens to every session and 7,523 once invoked, about $0.0004 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 2026-08-31.
Other commands, from other repositories
octo-schedule
Manage scheduled workflow jobs for the Claude Octopus scheduler.
discover
Run a full product discovery cycle — from outcome definition through opportunity mapping, prioritisation, and experiment design. Use when the team isn't sure what to build next, or before writing a PRD for a complex feature space.
linear-bulk
Apply the same change to many Linear issues at once. Dry-run by default.
prd-score
Score an existing PRD against the 100-point AI-optimization framework.
sdlc
Master orchestrator for the complete 5-phase SDLC pipeline. Runs all phases sequentially with approval gates and handoffs between each phase. Organizes artifacts into handoffs/, docs/, and projects/ directories.
todo
Quick capture tool for the Future table. For structured planning, use /design.