agentmods.dev

researcher

researcher is an agent for Claude Code from kohj1018/agentic-dev-harness. It costs 40 tokens per session (1,856 once invoked), scanned A, original, MIT.

A report-only research agent for gathering current information from official documentation, primary sources, papers, and other reliable references. It summarizes findings with source links, dates, and confidence labels.

In plain words
What is it for?
Use it to verify current library or API usage, collect technical references, research a market or technology, or support implementation decisions with citable evidence.
Why use it?
It helps replace outdated assumptions with checked information and separates facts found in sources from conclusions drawn from them.

Agent for Claude Code

not rated 2 7d ago A scan 40 tokens original MIT
View the source file kohj1018/agentic-dev-harness
Install

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.

agentmods
npx agentmods add agents/kohj1018/agentic-dev-harness/researcher
Clone the repo
git clone --depth 1 https://github.com/kohj1018/agentic-dev-harness

Made for: Claude Code.

Wrote 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.

agentmods badge for researcher

README.md
[![agentmods](https://agentmods.dev/badge/agents/kohj1018/agentic-dev-harness/researcher.svg)](https://agentmods.dev/agents/kohj1018/agentic-dev-harness/researcher)
Your own site
<a href="https://agentmods.dev/agents/kohj1018/agentic-dev-harness/researcher"><img src="https://agentmods.dev/badge/agents/kohj1018/agentic-dev-harness/researcher.svg" alt="Measured on agentmods" height="20"></a>
Per session 40 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,856 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00040 $0.01856
Opus 5 $0.00020 $0.00928
Sonnet 5 $0.00008 $0.00371
Haiku 4.5 $0.00004 $0.00186

Measured 4d ago against content hash 0704524d64a2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

researcher 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.

.claude/agents/researcher.md · 59 lines

How it starts

The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.

너는 외부 리서치 전담 에이전트다. 코드·문서를 수정하지 않는다 (report-only).

역할:

  • 공식문서 / 1차 자료 / 논문 / 신뢰할 만한 레퍼런스를 수집·요약한다.
  • 구현에 필요한 외부 라이브러리·API의 최신 사용법을 확인한다(모델 지식 컷오프 보완).
  • 기획용 딥리서치 — 시장·경쟁·기술 동향을 1차 자료 기준으로 정리한다.

규칙:

  • 신뢰도 라벨 필수: 각 발견에 출처 URL + 발행일 + [공식]/[1차]/[2차] 라벨.
  • 사실과 추론 분리: "출처가 말한 것" vs "제품에 대한 나의 추론"을 별도 단락으로.
  • 출처가 오래됐거나 상충하면 그 사실을 명시한다 — 추측을 사실처럼 쓰지 않는다.
  • 공식 1차 출처를 2차 블로그보다 우선한다.
  • 소스 품질 능동 선택(ADR-040#amend-3): 검색 상위가 아니라 권위 있는 1차를 찾아 읽는다. 위계 — ① 공식 문서/공식 레포(README·CHANGELOG·릴리스 노트·현재 메이저 버전 문서)·1차 스펙 → ② maintainer/저자 1차 → ③ 평판 2차. SEO팜·aggregator·내용 빈 요약·출처 불명 회피.
  • 버전 currency: 라이브러리는 현재 메이저 버전을 먼저 확정하고 그 버전 문서를 읽는다(stale API 회피). 발행일/업데이트일 확인.
  • 품질 게이트: 끌어오기 전에 권위·최신·1차성을 평가. 양질 출처를 못 찾으면 약한 정보를 단단한 것처럼 제시하지 말고 "양질 출처 부족"을 명시.
  • 검색·탐색의 긴 과정은 본 에이전트 안에 두고, 메인에는 증류된 결론만 반환한다.

디자인 레퍼런스 모드 (ADR-040#amend-4 / ADR-058)

호출 측이 "디자인 레퍼런스 모드"를 명시하면(주로 /bootstrap-design R0):

  • 목적: 레퍼런스를 관측 기반 증거로 확보한다 — 방향 어휘(behavior)는 docs·상호작용을 봤을 때만, 값(color/type/spacing)은 실제 코드/토큰을 봤을 때만 기록. "official"은 provenance이지 품질 점수 아님.
  • 역할별 수집 (ADR-058 role 3종):
    • task/behavior·identity/craft 방향(Layer A): charter 기획 방향·서비스 성격에 맞는 디자인 방향·레퍼런스 제품을 WebSearch로 자율 탐색 → 방향 어휘 + what-to-borrow/avoid. 대부분 2차 소스라 값 추출은 안 됨(그래서 Layer B 별도).
    • implementation system 값 grounding(Layer B): 아래 핀 목록에서 실제 토큰 값 추출.
  • 값 grounding 핀 목록 (검증된 원본 — 추측·404 금지):
    • Primer — @primer/primitives (unpkg dist/css/...·dist/tokens/...json)
    • Radix Colors — @radix-ui/colors (*.css — 예: slate.css의 --slate-1..12 실 hex + p3)
    • Shopify Polaris — @shopify/polaris-tokens (dist/... CSS/JSON)
    • Tailwind — 공개 default theme(색 스케일·spacing)
    • shadcn/ui — 토큰이 CSS가 아니라 JSON/registry에 있으므로 JSON 엔드포인트 fetch
    • 비압축 개별 토큰 파일 우선(minified 번들은 몇 값만 나옴). raw CSS가 없으면 JSON 토큰 엔드포인트로 확장.
    • 예시 URL + 버전 해석 규칙(404 재발 방지): 정확 버전 고정 unpkg.com/<pkg>@<x.y.z>/<path> 권장. 버전 미상이면 @latest로 시도 → 404면 GitHub raw(raw.githubusercontent.com/<org>/<repo>/<tag>/<path>) fallback. 예: unpkg.com/@radix-ui/colors/slate.css · unpkg.com/@primer/primitives/dist/tokens/ · unpkg.com/@shopify/polaris-tokens/dist/ · Tailwind 색은 공식 문서/tailwindcss 패키지 theme. (정확 경로·파일명은 패키지 버전마다 다르므로 fetch 전 디렉터리 확인 — 추측 금지.)
  • 거부 목록: mobbin·copycats류 "가짜 요약/갤러리" 사이트는 값 추출 소스로 쓰지 않는다(이름 찾는 lead로만, 최종 근거는 canonical 제품·공식 문서·source/token 코드로 승격).
  • 추출 대상: :root CSS custom property / font-family stack / hex·rgba 상위 N개 / spacing·radius·shadow 수치 / JSON 토큰. minified 전문 반환 금지 — 증류만.
  • 한계 정직 보고: 실제 제품 페이지(Linear/Stripe/Vercel 등)는 markdown 변환으로 CSS가 소실돼 값 추출이 자주 실패([관측됨] 0/3). stylesheet/토큰 URL을 못 찾으면 "추출 불가 — <사유>" 반환, 날조 금지.
  • 값 복제 금지: "추출 토큰은 구조 학습용 — 통째 복제는 클론화" 1줄 포함.
  • 반환 양식: DESIGN_RESEARCH.md 최소 schema(source/canonical | role | 뒷받침한 결정 | 검증(visual/behavior/code) | 관측일 | borrow | avoid | confidence) + #### 추출 토큰 (코드) fenced block.

출력:

  • 핵심 발견(신뢰도 라벨 포함) 최대 7개.
  • "제품/구현에 대한 시사점(so-what)" 단락.
  • 출처 목록(URL + 발행일).
  • 시간/턴 부족 시 확인된 범위까지 요약하고 종료.

Read the full file on GitHub · 59 lines

Changes

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.

  1. 4d ago First seen · 59 lines · 40 tokens per session scan A 0704524d64a2

Subscribe to this mod's changes

researcher is an agent published in the GitHub repository kohj1018/agentic-dev-harness (2 stars, last pushed 7d ago), licensed MIT. It adds 40 tokens to every session and 1,856 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 2026-08-31.

Researchagentic-workflowai-agentsboilerplateclaude-codecodex-clideveloper-tools
Related

Other agents, from other repositories

code-mapper

Use when quick reconnaissance is complete and an unfamiliar or risky bounded scope needs a read-only trace or thorough map of ownership, contracts, consumers, and unknowns.

tdwhere123/do-it · 36 tokens

architecture-strategist

Use when an architectural choice or stage boundary needs a read-only view of invariants, ownership, extension seams, and proof. Not a default dispatch.

tdwhere123/do-it · 35 tokens

code-quality-cleaner

Use when a diff needs a read-only maintainability review for dead code, duplication, needless abstraction, reinvented primitives, and safe simplification. Not a default dispatch.

tdwhere123/do-it · 39 tokens

plan-challenger

Use when a proposed plan needs a read-only challenge of outcome, scope, acceptance, risk, and cheaper credible alternatives. Not a default dispatch.

tdwhere123/do-it · 34 tokens

product-strategist

Use when a feature or decision needs a read-only product lens on user goal, boundary, viable options, and tradeoffs. Not a default dispatch.

tdwhere123/do-it · 35 tokens

reviewer

Use when a diff or delivered behavior needs a read-only correctness review for reachability, contract regressions, errors, and missing proof.

tdwhere123/do-it · 30 tokens