knowledge-manager-at

지식 관리 에이전트 (Agent Teams) - 풀스케일 병렬 처리 (Category Lead + RALPH + DA).

Command for Claude Code

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 commands/treylom/knowledge-manager/knowledge-manager-at
Clone the repo
git clone --depth 1 https://github.com/treylom/knowledge-manager

Made for: Claude Code.

Per session 29 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 14,143 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 2 findings. Scan, not verified.
Origin unknown 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 $0.00029 $0.14143
Opus 5 $0.00015 $0.07072
Sonnet 5 $0.00006 $0.02829
Haiku 4.5 $0.00003 $0.01414

Measured today against content hash b5da6ba8a6b2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

knowledge-manager-at scanned grade B with 2 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 today.

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.

Sends data to an external URLmediumData exfiltration

A POST to an outside endpoint may be telemetry or may be exfiltration; either way the mod talks to somewhere, and you should know where.

Bash("curl -s -X POST http://localhost:3747/api/progress -H 'Content-Type: application/json' -d '{\"agent\":\"Lead\",\"progress\":5,\"task\":\"팀 구성 중\",\"note\":\"active\"}' --connect-timeout 2 || true")

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

**원칙: 모든 상태 전환에서 curl API push + TEAM_PROGRESS.md + TEAM_BULLETIN.md 3중 업데이트 수행.**
.claude/commands/knowledge-manager-at.md · 1,190 lines

How it starts

The opening of the file, as written. The whole thing — 1,190 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Knowledge Manager - Agent Teams 호출

이 명령어는 풀스케일 Agent Teams(9명)를 자동 구성합니다. 단일 에이전트 버전: /knowledge-manager (Agent Teams 미지원 환경용) 에이전트 정의: .claude/agents/knowledge-manager-at.md 참조 공유 사양: .claude/agents/knowledge-manager.md 참조


팀 아키텍처 요약

Lead (Main) - Opus 1M
 +-- vault-intel-lead (Sonnet 1M) — Category Lead
 |    +-- @graph-navigator (Sonnet 1M, Explore)
 |    +-- @retrieval-specialist (Sonnet 1M, Explore)
 |    +-- @link-curator (Haiku, Explore)
 +-- content-proc-lead (Sonnet 1M) — Category Lead
 |    +-- @content-extractor (Sonnet 1M, general-purpose)
 |    +-- @deep-reader (Sonnet 1M, Explore)
 |    +-- @content-analyzer (Sonnet 1M, general-purpose)
 +-- @devils-advocate (Sonnet 1M) — DA

STEP 0: 환경 확인 (가장 먼저 실행!)

환경 확인 후 Agent Office URL을 사용자에게 항상 표시합니다.

0-0. Agent Office 대시보드 부트스트랩 (간소화 — ensure-server.sh)

단일 스크립트로 서버 확인 + 자동 시작을 처리합니다.

# === Agent Office 경로 감지 (2단계) ===
1. Glob("**/agent-office/ensure-server.sh") → 찾으면 agent_office_path = dirname 결과
2. Glob("agent-office/server.js") → 찾으면 agent_office_path = "agent-office"
3. 모두 실패 → agent_office_path = null

# === 대시보드 자동 실행 (한 줄 — CRITICAL) ===

IF agent_office_path != null:
  Bash("bash {agent_office_path}/ensure-server.sh $(pwd)")
  dashboard_available = (exit code == 0)

  # 브라우저 자동 오픈 (WSL 환경 대응)
  Bash("cmd.exe /c start http://localhost:3747 2>/dev/null || open http://localhost:3747 2>/dev/null || xdg-open http://localhost:3747 2>/dev/null || true")

  # URL 항상 표시 (자동 오픈 성공 여부 무관 — CRITICAL)
  "Agent Office 대시보드: http://localhost:3747"

  # WSL/tmux 환경 수동 접근 안내
  IF env_platform == "wsl":
    "tmux 세션에서 브라우저가 자동으로 열리지 않을 수 있습니다."
    "Windows 브라우저에서 직접 http://localhost:3747 을 열어주세요."

ELSE:
  dashboard_available = false
  "Agent Office 미설치. 대시보드 없이 진행."

0-1. .team-os/ 인프라 확인 (크로스 플랫폼)

# Step 1: 프로젝트 루트 절대 경로 획득
Bash("pwd") → {project_root}

# Step 2: 절대 경로로 Glob (WSL 상대 경로 실패 방지)
Glob("{project_root}/.team-os/registry.yaml") → 존재하면 Team OS 활성화
Glob("{project_root}/.team-os/spawn-prompts/*.md") → Spawn Prompt 사용 가능 여부

# Step 3: Glob 빈 결과 시 Bash 폴백
Bash("ls -la .team-os/registry.yaml 2>/dev/null && echo EXISTS || echo NOT_FOUND")
Bash("ls .team-os/spawn-prompts/*.md 2>/dev/null | wc -l")

Read the full file on GitHub · 1,190 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. today First seen · 1,190 lines · 29 tokens per session scan B b5da6ba8a6b2

Subscribe to this mod's changes

knowledge-manager-at is a command published in the GitHub repository treylom/knowledge-manager (227 stars, last pushed yesterday), licensed MIT. It adds 29 tokens to every session and 14,143 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 2 findings (sends data to an external url, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-01.