microsoft/mcp is a repository of Microsoft’s Model Context Protocol server implementations, which let AI applications connect to data sources and tools through a shared client-server interface. It supports Microsoft MCP server contributors and the catalogue entries are add-ons for working with these servers.
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 microsoft/mcp --skill harden-tool-url-inputsgit clone --depth 1 https://github.com/microsoft/mcpWrote 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/microsoft/mcp/harden-tool-url-inputs)<a href="https://agentmods.dev/skills/microsoft/mcp/harden-tool-url-inputs"><img src="https://agentmods.dev/badge/skills/microsoft/mcp/harden-tool-url-inputs/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/microsoft/mcp/harden-tool-url-inputs"><img src="https://agentmods.dev/badge/skills/microsoft/mcp/harden-tool-url-inputs.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 | $0.00077 | $0.05786 |
| Opus 5.5 | $0.00031 | $0.02314 |
| Sonnet 5.5 | $0.00015 | $0.01157 |
| Haiku 4.5 | $0.00008 | $0.00579 |
Grade A, and why
harden-tool-url-inputs 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 12d 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 — 355 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Harden Tool URL Inputs
Audit and update every requested directory so MCP-influenced values cannot hijack a URL, URI, host, or client endpoint. Use IAzureService forwarding methods or the host's injectable IEndpointValidator at the final URL boundary, with the configured Azure cloud and the narrowest appropriate allow-list.
This is a scoped implementation skill, not a report-only audit. Make the fixes and add tests unless the user explicitly requests analysis only.
Required Input and Scope
- Accept one or many repository-relative directories under
tools/. - Resolve every path and confirm it is inside the repository's
tools/directory. Reject files or directories outside that boundary. - If no directory is supplied, ask the user for at least one. If a supplied path does not exist or its intended scope is ambiguous, ask before editing.
- Treat each supplied directory as an exhaustive scope. Do not stop after finding the first vulnerable URL flow.
- Do not modify other tool directories. Shared changes to
core/Microsoft.Mcp.Core/src/Helpers/EndpointValidator.AllowLists.csand its tests are allowed when a scoped tool needs a new Azure service allow-list. - Preserve unrelated worktree changes. Read the repository and nearest scoped instructions before editing.
Pattern Sources
Before implementing, inspect the current versions of:
core/Microsoft.Mcp.Core/src/Helpers/IEndpointValidator.cscore/Microsoft.Mcp.Core/src/Helpers/EndpointValidator.cscore/Microsoft.Mcp.Core/src/Helpers/EndpointValidator.AllowLists.cscore/Azure.Mcp.Core/src/Services/Azure/IAzureService.cscore/Microsoft.Mcp.Core/tests/Microsoft.Mcp.Core.Tests/Helpers/EndpointValidatorTests.cscore/Azure.Mcp.Core/tests/Azure.Mcp.Tests/Helpers/AzureServiceTestHelpers.cs
When wiring dependencies or testing namespace overrides, also inspect SsrfProtectionPolicy.cs,
EndpointValidationServiceCollectionExtensions.cs, ICommandContextAccessor.cs, and
EndpointValidationRegistrationTests.cs in their corresponding core projects. Read the
executing-command-context guidance in CONTRIBUTING.md and transport boundaries in
docs/design/HttpClientFactory.md.
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.
- 12d ago Changed · +83 lines · +6 tokens per session 38e57d6ca8e1
- 22d ago First seen · 272 lines · 71 tokens per session scan A 9d3318c6fca5
harden-tool-url-inputs is a skill published in the GitHub repository microsoft/mcp (3,800 stars, last pushed yesterday), licensed MIT. It adds 77 tokens to every session and 5,786 once invoked, about $0.0003 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 2026-10-03.
Other skills, from other repositories
research-engineer
An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.
tika-eval-compare
Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.
jetson-validate-image
Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.
atmos-validation
Validate Atmos projects, components, arbitrary JSON Schema inputs, EditorConfig, and GitHub Actions; use affected-file selection and native CI annotations.
skill-benchmark
Benchmark AI skill effectiveness by measuring implementation quality against legacy constraints.