MCP Composer is a FastMCP server that combines tools from multiple Model Context Protocol servers behind one endpoint. It is for exposing member servers that use HTTP, SSE, stdio, OpenAPI, or GraphQL through a single configured tool surface. The catalogue instruction supports working with this composer.
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.
npx skills add IBM/mcp-composer --skill bob-skillsgit clone --depth 1 https://github.com/IBM/mcp-composerWrote 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/skills/ibm/mcp-composer/bob-skills)<a href="https://agentmods.dev/skills/ibm/mcp-composer/bob-skills"><img src="https://agentmods.dev/badge/skills/ibm/mcp-composer/bob-skills/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ibm/mcp-composer/bob-skills"><img src="https://agentmods.dev/badge/skills/ibm/mcp-composer/bob-skills.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00025 | $0.02693 |
| Opus 5.5 | $0.00010 | $0.01077 |
| Sonnet 5.5 | $0.00005 | $0.00539 |
| Haiku 4.5 | $0.00003 | $0.00269 |
Grade A, and why
tool-audit 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 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.
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 — 324 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Tool Definition Audit Skill
This skill performs a systematic audit of MCP tool definitions to identify potential sources of confusion for agent decision-making.
Parameters Required
The user will provide the JSON file path in their request. If not provided, ask for:
- tools_json_path: Path to the JSON file containing tool definitions (e.g.,
tools.json)
Expected JSON Format
The skill expects a JSON structure with:
{
"active": {
"server-name": [
{"name": "tool_name", "description": "Tool description"}
]
},
"deactivated": ["server-list"]
}
Note: Some servers may have ERROR entries (authentication/connectivity failures) that should be excluded from real tool counts.
Audit Process
Count active servers and tools
active_servers = len(data['active']) total = 0 for server, tools in data['active'].items(): count = len(tools) total += count print(f'{server}: {count} tools') print(f'TOTAL: {total} tools') print(f'Active servers: {active_servers}')
Count deactivated servers
deactivated = data.get('deactivated', []) deactivated_count = len(deactivated) print(f'Deactivated servers: {deactivated_count}') if deactivated_count > 0: print(f'Deactivated list: {deactivated}')
- **NEVER estimate** based on file size, line count, or manual counting
- **ALWAYS use Python** to count tools and servers programmatically
- Count both active AND deactivated servers
- Iterate through each server in "active" section
- Count tools per server: `len(tools_array)`
- Count deactivated servers: `len(data.get('deactivated', []))`
- Identify ERROR tools (name starts with "ERROR" or description contains "HTTP Error")
- Calculate: Real tools = Total entries - ERROR tools
- Verify total by summing all server counts
- Extract all tool names and descriptions from active servers
- Create inventory with server/product associations
- Calculate statistics:
- Total active servers: `len(data['active'])`
- Total deactivated servers: `len(data.get('deactivated', []))`
- List deactivated server names if any exist
- Total real tools (excluding ERROR placeholders)
- Tools per server breakdown with accurate percentages
- Description length distribution
- Quality score distribution
- **VALIDATION**: Before proceeding, verify that:
- Sum of all server tool counts equals reported total
- Percentages add up to 100%
- No server has 0 tools unless it's truly empty or has ERROR
- Deactivated server count matches the actual list length
</Step>
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- today First seen · 324 lines · 25 tokens per session scan A c6c07c5305ac
tool-audit is a skill published in the GitHub repository IBM/mcp-composer (22 stars, last pushed today), licensed Apache-2.0. It adds 25 tokens to every session and 2,693 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-11-14.
tool-audit compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| tool-auditIBM/mcp-composer | Skill | 22 | today | Apache-2.0 | A | 25 | Claude Code, Codex |
| imagegenopenai/codex | Skill | 135k | today | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 100k | today | Apache-2.0 | A | 19 | Claude Code |
| browser-automationopenclaw/openclaw | Skill | 394k | today | MIT | A | 36 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 80k | 7d ago | MIT | A | 200 | Claude Code, Codex |
| next-cache-components-adoptionvercel/next.js | Skill | 144k | today | MIT | A | 95 | Claude Code |
| chat-pet-sprite-creationmicrosoft/vscode | Skill | 195k | today | MIT | A | 53 | Claude Code, Codex |