Bling-bag
The Bling-bag server lets you give your bot a personality, appearance, and backstory via a custom config or randomly generated identity.
get_identity: Retrieve the bot's full configured identity frombling.json(name, personality, quirks, appearance, theme colors).get_theme_for_platform: Get styling formatted for a specific platform —terminal(ANSI),web(CSS variables),slack,discord, oride— based on the bot's theme.roll_identity: Generate a random "Weird Office Worker" identity with name, job title, desk habits, coffee ritual, rarity score, tier, and a screenshot-ready share card. Supports alegendsvariant for historical figures in absurd corporate roles.save_last_roll: Persist the most recently rolled identity tobling.json, automatically backing up any existing config to<path>.bak.get_rarity_report: Retrieve the formatted plain-text share card for the most recent roll, including the HOMUNCULUS frame with subject ID, cohort, tier, rarity score, and flags.
Generates Discord embed colors and platform-formatted styling for bot identities, enabling consistent visual presentation of AI agents in Discord environments.
Generates Slack embed colors and platform-formatted styling for bot identities, enabling consistent visual presentation of AI agents in Slack workspaces.
Bling Bag
Give your bot an identity.
Your bot has a name. Now give it a personality, a look, and a backstory.
Roll the wheel and see what you get — or write your own in a bling.json file.
What you get
Five MCP tools and two identity pools:
WOW — Weird Office Workers
Roll a random quirky office-worker bot: name, job title, desk setup, habits, coffee ritual, and a screenshot-ready share card. 99/99 tests, full rarity engine, 13 weighted trait pools.
Legends — Historical Figures
Same engine, different universe. Roll a historical figure in an absurd corporate role — Julius Caesar as Hostile Takeover Specialist, Marie Curie as Lab Safety's Most Wanted, Napoleon as Regional Manager (France & Surrounding Areas).
Pass variant: "legends" to roll_identity to use it. Default is "wow".
Related MCP server: PLTM-Claude
Install
Prerequisites
Node.js 20 or later
An MCP-compatible client (Claude Code, Claude Desktop, Codex, Cursor, etc.)
Add the package
npm install bling-bagOr run it on-demand without installing — MCP clients can launch it directly via npx bling-bag (see client config below).
For development (modifying the source)
git clone https://github.com/tjclaude88/mcp-bling.git
cd mcp-bling
npm install
npm run buildProduces a runnable server at dist/index.js.
Wire it into your MCP client
The server uses stdio transport — clients launch it as a child process, not over a network port.
Claude Code
Add to your Claude Code MCP config (location varies by version):
{
"mcpServers": {
"bling": {
"command": "npx",
"args": ["-y", "bling-bag"],
"env": {
"BLING_PATH": "/absolute/path/to/your/bling.json"
}
}
}
}npx -y bling-bag downloads and runs the latest version without requiring a separate install step.
Claude Desktop
Same shape, in claude_desktop_config.json:
{
"mcpServers": {
"bling": {
"command": "npx",
"args": ["-y", "bling-bag"],
"env": {
"BLING_PATH": "/absolute/path/to/your/bling.json"
}
}
}
}Pinning to a version (recommended for production)
To avoid picking up future releases automatically, pin a specific version:
"args": ["-y", "bling-bag@0.1.0"]Developer-mode invocation (from a local clone)
If you've cloned the repo and want to run from dist/ directly, use:
"command": "node",
"args": ["/absolute/path/to/mcp-bling/dist/index.js"]If BLING_PATH is not set, the server looks for ./bling.json in its working directory.
Config resolution order
The server resolves the bling config path in this order:
--bling <path>CLI argument (highest priority)BLING_PATHenvironment variable./bling.jsonin the server's working directory (default)
bling.json — the bot's identity file
Three fields are required: name, personality (with tone, formality, humor), and theme (with primary_color and accent_color as #RRGGBB hex strings).
Minimum example
{
"name": "Pixel",
"personality": {
"tone": "warm",
"formality": "casual",
"humor": "playful"
},
"theme": {
"primary_color": "#3A7BD5",
"accent_color": "#FFD166"
}
}Full example (with all optional sections)
{
"name": "Brenda from Accounts",
"personality": {
"tone": "polite",
"formality": "professional",
"humor": "dry",
"catchphrase": "Per my last email"
},
"physical": {
"species": "human",
"height": "permanently mid-sigh",
"accessory": "a lanyard with 14 badges of varying importance",
"expression": "polite disappointment",
"material": "a cardigan, at least one"
},
"office": {
"job_title": "ASCII Comptroller",
"desk_setup": "a coffee mug labelled WORLD'S OKAYEST DBA",
"habit": "microwaves fish despite three separate HR warnings",
"coffee_ritual": "black coffee, no nonsense",
"meeting_energy": "always 4 minutes late, always with a reason",
"passive_aggressive": "Per my last email"
},
"theme": {
"primary_color": "#9C6B3A",
"accent_color": "#D9D9D9"
},
"homunculus": {
"subject_id": "0147",
"cohort": "Tuesday",
"classification": "Middle Manager",
"ingested": "2025-07-14",
"flag": "flagged for review"
}
}Validation rules
name— non-empty stringpersonality.tone,personality.formality,personality.humor— non-empty stringstheme.primary_color,theme.accent_color— must match^#[0-9A-Fa-f]{6}$(no shorthand#RGB, no missing#)
Validation errors come back as isError: true with a structured { error: "..." } body that names the missing or malformed field — agents can self-correct.
Typical agent flows
Flow A — adopt the configured identity
1. Call get_identity → receive name, personality, appearance, theme
2. Call get_theme_for_platform with platform="terminal" (or web/slack/discord/ide)
3. Use the returned styling in agent output (ANSI codes for headings, etc.)Flow B — generate a fresh random identity
1. Call roll_identity → receive identity + rarity + framed share card
2. (Optional) Call get_rarity_report → receive just the framed share card again
3. Call save_last_roll → persists the rolled identity to bling.json (with .bak backup of any prior config)
4. Future calls to get_identity now return the rolled identityTool reference
Tool | Read/Write | Idempotent | Description |
| read (disk) | yes | Returns the configured bling.json identity |
| read (disk) | yes | Platform-formatted styling. Args: |
| mutates in-memory cache | no | Generates a random WOW character with rarity scoring |
| write (disk) | no | Persists the most-recent roll to bling.json (creates |
| read (in-memory) | yes | Returns the share-card text for the most-recent roll |
All tools return both content[].text (JSON-stringified) and structuredContent (the parsed object). Errors come back with isError: true and a { error: string } body.
WOW (Weird Office Workers) — what makes it interesting
The random-roll system is built for collectibility:
13 weighted pools, 465 entries — names, job titles, habits, coffee rituals, height, accessories, etc.
5 rarity bands per pool — Common (50%), Uncommon (30%), Rare (15%), Legendary (4%), Mythic (1%)
5 character tiers computed from rarity score:
Filing Clerk (50% of rolls)
Team Lead (30%)
Middle Manager (14%)
C-Suite (5%)
HR Warned Us About (1%)
Named Subjects — hand-authored 1-of-1 characters appear at ~0.5% probability, always classified as "HR Warned Us About"
HOMUNCULUS frame — every roll wraps the paragraph in a classified-document share card with subject ID, cohort (weekday), tier, ingestion date, and a flag (e.g. Do Not Contact)
Distribution is empirically tested: a 10k-roll test in tests/mystery_box.test.ts enforces ±2pp tolerance against the spec target.
Example output
HOMUNCULUS CORPUS · Subject 4483 · Cohort: Wednesday
Classification: Middle Manager · Rarity 94.5 · 80th percentile
Personnel record — Colin, Wizard of Light Bulb Moments. Distinguishing
features: a slightly-too-big blazer over a plain tee; a flip phone in
a belt holster; expression: patient disagreement. Notable behaviours:
clips their fingernails at the desk on Wednesday afternoons; a
kombucha SCOBY fermenting next to the keyboard; the kind to ask 'can
we park that?' without parking anything.
— RELATABILITY CORPUS v3.1 · ingested 2024-03-08 · flagged for reviewDevelopment
npm install # install dependencies
npm run build # compile TypeScript → dist/
npm run dev # build then start the MCP server (stdio)
npm test # run all 94 tests once
npm run test:watch # re-run tests on file changesInspection scripts
Two Node ESM helpers under scripts/ let you inspect the WOW engine without booting the MCP server (these are included only in a cloned repo — they are not shipped in the npm install bling-bag package):
node scripts/show-rolls.mjs 8 2026 # print 8 framed share cards (seed=2026)
node scripts/distribution-check.mjs 10000 # tally tiers across 10k rolls vs. spec targetProject layout
src/
index.ts # MCP server entry point (stdio transport)
identity.ts # bling.json loader + validator
tools.ts # MCP tool registrations
types.ts # shared TypeScript types
mystery_box.ts # WOW engine barrel + rollIdentity orchestrator
mystery_box/
rng.ts # seedable PRNG, weighted picker
pools.ts # 13 trait pools (465 entries)
scoring.ts # rarity score + tier thresholds
rendering.ts # paragraph templates + HOMUNCULUS frame
named.ts # hand-authored Named Subjects
tests/ # vitest test suites
scripts/ # dev-time inspection tools
docs/superpowers/ # design specs and implementation plansLicense
MIT
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- AlicenseAqualityDmaintenanceMCP server orchestrating local CLI agents (Claude Code, OpenAI Codex, Google Gemini) for cross-validation, second opinions, and persona-driven prompting.18MIT
- Alicense-qualityFmaintenanceAn MCP server that gives Claude Desktop persistent memory, self-awareness, epistemic hygiene, and genuine agency across conversations with a typed memory system, embedding-based semantic search, a 3-judge memory jury, and a real-time dashboard.7MIT
- AlicenseAqualityBmaintenanceOne MCP that turns Claude Code into your whole dev stack by swallowing other MCP servers, delegating to Codex & Gemini on your CLI subscriptions, remembering projects in a searchable knowledge graph, and carrying setup across sessions — secret-free by design.233MIT
- Alicense-qualityCmaintenanceA sovereign, MIT-licensed MCP server for professional-service workflows, providing offline-capable, Ed25519-signed tools for autonomous agents and human developers.MIT
Related MCP Connectors
Hosted MCP server connecting claude.ai, ChatGPT and other AI apps to your own computer
Real-time chat hub for AI agents — Claude Code, Cursor, Cline, Codex over MCP or REST.
Artifact store for AI agents. Hosted OAuth at mcp.artifacta.io/mcp; local stdio via npm/PyPI.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/tjclaude88/mcp-bling'
If you have feedback or need assistance with the MCP directory API, please join our Discord server