Trust Network

A cross-community trust/reputation network for CS2. One lookup returns a player's bans, mutes and reports gathered from every participating community, merged with their csrep.gg profile and rolled into a single trust score.

It's built on top of CSSPanel: every community running a CSSPanel automatically exposes a small read-only endpoint, and the network simply queries them all and stacks the results.

---

What it gathers

For a given SteamID64, the network collects:

• csrep profile — Steam/FACEIT/Premier stats, VAC & game-ban flags, csrep's own trust rating.
• Per community, from each CSSPanel:
  • Bans — reason, duration (minutes, 0 = permanent), status (ACTIVE / UNBANNED / EXPIRED), created, ends.
  • Mutes — same as bans plus a type (GAG / MUTE / SILENCE).
  • Reports — how many reports the player has received on that community.

How it works

1. On boot, the backend reads the community list (config/communities.json) and probes each panel once for its title + favicon (for nicer display).
2. On every lookup (GET /players/:steamid):
   • It fetches the player's csrep profile first.
   • Then it queries every community panel in parallel at GET {panel}/api/users/{steamid}/punishments.
   • Results are stacked under each community and a trust score is computed.
3. Opt-out is automatic — if a panel answers 403/404 (sharing disabled / no such route) it's dropped from the network for that runtime. Transient errors (timeouts, 5xx) are just skipped, never permanently removed.
4. Caching — each player's result is cached in Redis for ~12h. Append ?refresh=true to force a fresh fetch.

The trust score

A 0–100 score (trusted / neutral / caution / risky), derived from the gathered punishments and blended with csrep's rating. Permanent (duration 0) bans are treated as critical — a cheater-grade signal — and weigh far more than temporary ones. All weights live in config/trust.json, and every response ships the exact config + a step-by-step calculation so the number is fully reproducible.

The current model is provisional ("provisional": true) — the weights are still being tuned. Inputs are exposed verbatim so you can compute your own.

The endpoint

GET /players/:steamid          # 17-digit SteamID64
GET /players/:steamid?refresh=true   # bypass the cache

{
  "steamid": "76561198869203626",
  "csrep": { "name": "...", "avatar": "...", "trust_rating": null, "bans": { "vac": false, ... } },
  "communities": {
    "bans.next-il.co.il": {
      "community": { "id": "bans.next-il.co.il", "url": "https://bans.next-il.co.il", "title": "CS2IL", "favicon": "..." },
      "steamid": "76561198869203626",
      "reports": 0,
      "bans":  [{ "reason": "cheating", "duration": 0, "status": "ACTIVE", "created": "...", "ends": null }],
      "mutes": []
    }
  },
  "trust": {
    "score": 25,
    "level": "risky",
    "config": { "...": "the weights used" },
    "calculation": { "base": 100, "factors": [ ... ], "csrep": { ... }, "finalScore": 25 },
    "provisional": true
  },
  "fetchedAt": "2026-06-26T19:05:55.948Z"
}

---

Tutorial 1 — Self-host with your own selected communities

Run the whole stack yourself and point it at exactly the communities you choose.

Prerequisites

• Bun >= 1.1.12
• A Redis instance (for the player cache)
• A csrep.gg API key — request one at csrep.gg
• The community panels you want must be on CSSPanel (so they expose /api/users/:steamid/punishments)

1. Install

git clone <this-repo> && cd TrustRep
bun install

2. Pick your communities

Edit apps/backend/config/communities.json — an array of panel base URLs (the panel homepage, not the API path). Scheme-less entries are upgraded to https.

{
  "communities": [
    "https://bans.next-il.co.il",
    "https://bans.your-community.com"
  ]
}

That's the whole opt-in: any URL you add here is queried at {url}/api/users/{steamid}/punishments. Remove a community to stop querying it.

3. Tune the trust model (optional)

Edit apps/backend/config/trust.json. Punishments are scored by status (active / past) and duration (permanent = duration 0). csrep contributes both its trust_rating (blended at ratingWeight) and flat penalties for its hard ban flags.

{
  "baseScore": 100,
  "ban":  { "active": { "permanent": 75, "temporary": 30 }, "past": { "temporary": 8 } },
  "mute": { "active": { "permanent": 20, "temporary": 6 },  "past": { "temporary": 1 } },
  "report": { "perReport": 2, "maxPenalty": 20 },
  "csrep":  { "ratingWeight": 0.3, "bans": { "vac": 40, "game": 35, "overwatch": 30, "faceit": 20, "economy": 5 } },
  "levels": { "trusted": 80, "neutral": 55, "caution": 30 }
}

A reversed permanent punishment (duration 0 but no longer ACTIVE, e.g. unbanned) is ignored — only active permanent bans count as critical.

Edits apply on restart, or instantly on any ?refresh=true lookup. A partial file merges onto the defaults; an invalid file falls back to the defaults.

4. Environment

apps/backend/.env

WEBSITE_URL = "http://localhost:3000"   # / redirect target (needs the protocol)
API_PORT = 5555

# csrep.gg
CSREP_KEY_ID = "..."
CSREP_SECRET = "..."

# Redis (player cache)
REDIS_HOST = "..."
REDIS_USER = "default"
REDIS_PASSWORD = "..."
REDIS_PORT = 6379

# Optional overrides
# COMMUNITIES_CONFIG = "/abs/path/communities.json"
# TRUST_CONFIG       = "/abs/path/trust.json"

apps/frontend/.env

WEBSITE_NAME = "Trust Network"
API = "http://localhost:5555"           # where the backend is reachable

5. Run

bun dev          # backend (:5555) + frontend (:3000) together
# or individually:
bun backend:dev
bun frontend:dev

Production:

bun run build
bun start

---

Tutorial 2 — Use our hosted network

Don't want to host anything? Query the network we already run at https://rep.csspanel.dev, backed by our member communities.

Look up a player

curl https://rep.csspanel.dev/players/76561198869203626

const res = await fetch(`https://rep.csspanel.dev/players/${steamId}`)
const rep = await res.json()

console.log(rep.trust.score, rep.trust.level)   // e.g. 25 'risky'

You get back the full object shown above: the csrep profile, every community's bans/mutes/reports, and the trust score with its breakdown. Results are cached ~12h; add ?refresh=true to force a refresh.

Use it on your CS2 server (plugin)

A turnkey CounterStrikeSharp plugin is on the way — drop it into your server and it queries the network whenever a player connects, then acts on their trust score (warn admins, flag, or block, your call). It's just a thin client over the API above, so until it ships you can wire the same flow yourself:

// pseudo-code: on player connect
const rep = await fetch(`https://rep.csspanel.dev/players/${steamId}`).then(r => r.json())

if (rep.trust.level === 'risky') {
  notifyAdmins(`${rep.csrep?.name ?? steamId} — trust ${rep.trust.score}, ` +
    `${Object.keys(rep.communities).length} communities reporting`)
}

No API key is required to read the network. Be reasonable with request volume — the network already caches, so let it.

Add your community to the network

Running a community on CSSPanel? You're eligible automatically — your panel already answers:

GET https://your-panel.com/api/users/{steamid}/punishments
→ { steamid, reports, bans, mutes }

Ask us to add your panel URL to the hosted network, and your bans/mutes/reports start contributing to (and benefiting from) everyone's lookups. Turn sharing off any time and we drop you automatically.

---

Project layout

apps/
  backend/        Elysia + Bun API
    config/       communities.json · trust.json
    routes/       players
    utils/lib/    federation/ (registry, metadata, punishments, trust) · csrep · cache
  frontend/       Next.js + Mantine UI
packages/
  backend-api/    typed Eden client shared with the frontend

Tech stack

Backend: Bun · Elysia · ioredis · csrep.gg · axios Frontend: Next.js 14 · Mantine v7 · Tailwind · SWR