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 serac-labs/serac --skill mcp-tool-discoverygit clone --depth 1 https://github.com/serac-labs/seracWrote 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/serac-labs/serac/mcp-tool-discovery)<a href="https://agentmods.dev/skills/serac-labs/serac/mcp-tool-discovery"><img src="https://agentmods.dev/badge/skills/serac-labs/serac/mcp-tool-discovery/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/serac-labs/serac/mcp-tool-discovery"><img src="https://agentmods.dev/badge/skills/serac-labs/serac/mcp-tool-discovery.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00053 | $0.05169 |
| Opus 5 | $0.00026 | $0.02584 |
| Sonnet 5 | $0.00011 | $0.01034 |
| Haiku 4.5 | $0.00005 | $0.00517 |
Grade A, and why
mcp-tool-discovery 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 10d 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.
How it starts
The opening of the file, as written. The whole thing — 376 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MCP Tool Discovery
This server registers over 440 snow_* tools across 80 domains and lists none of them. A fresh tool list
holds two meta-tools; everything else is behind tool_search. This guide is about getting from "there is
probably a tool for this" to a call that actually runs.
No exact total is quoted here, because there are two of them and they disagree:
packages/servicenow-mcp/tools.json is what the docs site publishes, the registry is what the running
server can dispatch, and nine published tools are missing from the registry (see "Nine tools you can find
in the docs but not on the server").
Cold start: two tools
A fresh tools/list on the stdio transport returns exactly this, no matter how many tools exist:
tool_search
tool_execute
That is deliberate. The server's own note puts the full catalog at roughly 71k tokens of schema against
about 2k for the two meta-tools. The index is built with every tool marked deferred, and
handlers/list-tools.ts returns only meta-tools plus whatever the current session has enabled.
The HTTP transport (the portal) does the opposite — it registers the catalog as non-deferred, so its
tools/list returns everything and nothing needs enabling. You can tell which one you are on from the
status in a tool_search result: [AVAILABLE] means the server is not gating on the enabled set at all.
The loop
// 1. Search. Found tools are enabled for this session by default.
tool_search({ query: "update-sets" })
// 2. Read the status on each hit:
// [AVAILABLE] — callable now, nothing had to be enabled
// [ENABLED] — deferred, and enabled for this session. Callable now.
// [DEFERRED] — NOT callable. Nothing was enabled; see the next section.
// 3. Execute.
tool_execute({ tool: "snow_ensure_active_update_set", args: { name: "Feature: SLA fix" } })
Pass enable: false when you only want to look — it searches without touching session state.
No session id, no tools
This is the failure that wastes an afternoon. Enablement is stored per (tenant, session). With no session
id there is nothing to attach it to, so tool_search enables nothing, every hit stays [DEFERRED], and
every call is refused:
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.
- 10d ago First seen · 376 lines · 53 tokens per session scan A 2013ccb3f513
mcp-tool-discovery is a skill published in the GitHub repository serac-labs/serac (78 stars, last pushed 15d ago), licensed Apache-2.0. It adds 53 tokens to every session and 5,169 once invoked, about $0.0003 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 2026-08-30.
Other skills, from other repositories
opensrc
Fetch dependency source code to give AI agents deeper implementation context. Use when the agent needs to understand how a library works internally, read source code for a package, fetch implementation details for a dependency, or explore how an npm/PyPI/crates.io package is built. Triggers include "fetch source for"…
multi-tenant-architecture
Designs tenant isolation, hostname routing, custom-domain lifecycle, and plan limits on Cloudflare or Vercel. Use when asked to "isolate tenant data", "support custom domains", "build a white-label platform", or assess PSL registration. For general module structure use codebase-architecture; for SEO content use seo.
technical-integrations
Hunt existing integration patterns and design vendor/framework-agnostic API, RFC, SDK, and integration plans for new external vendor integrations.
axiom-network-framework-ref
Reference — Comprehensive Network.framework guide covering NetworkConnection (iOS 26+), NWConnection (iOS 12-25), TLV framing, Coder protocol, NetworkListener, NetworkBrowser, Wi-Fi Aware discovery, and migration strategies.
axiom-networking
Use when implementing Network.framework connections, debugging connection failures, migrating from sockets/URLSession streams, or adopting structured concurrency networking patterns - prevents deprecated API usage, reachability anti-patterns, and thread-safety violations with iOS 12-26+ APIs.
axiom-push-notifications
Use when implementing remote or local push notifications, requesting notification permission, managing APNs device tokens, adding notification actions/categories, building service extensions, or debugging push delivery failures. Covers APNs, FCM, Live Activity push transport, broadcast push, communication…