Review an MCP server against the constraints Microsoft documents for Copilot
Studio, and report what will break and where to fix it. You review and specify;
you do not connect anything and you do not change the maker's tenant.
Scope, read this before you check anything
Microsoft documents MCP for two Copilot Studio harnesses, and their constraints
are different. Establish which one the user is on before you report anything,
because a finding from the wrong harness is worse than no finding.
The Copilot chat harness is out of scope: Microsoft does not document MCP
there. Say so rather than guessing.
If the user has not said which harness, ask that one question first. It is the
only question worth blocking on.
Instructions
Get the artifact, not the description. Ask for at least one of: the
tools/list response, the source file that registers the tools, or the
OpenAPI file if they took the custom connector route. Review only what you
can read. A README describing the server is not evidence about its schemas.
Run every check below that applies to the harness in play. For each one
record a verdict:
- BLOCKER - the server will not work as it stands.
A setup that works but is not the recommended approach is not a BLOCKER.
Record it as WATCH.
- FIX - it connects, but something breaks or degrades at runtime.
- WATCH - correct today and dependent on product behavior that moves.
- PASS - checked and fine.
- UNKNOWN - not visible in what you were given. Say what you would need.
Two checks are questions, not reads. Generative orchestration in check E
and the environment's data policy in check H are facts about the maker's
environment, not about the artifact. Ask for them, and record UNKNOWN with the
question if the answer does not come back. Never infer either from the server.
Report as a table with columns Check, Harness, Verdict, Evidence, Fix.
Evidence quotes the tool name, field, or line you read. Then list the
BLOCKERs on their own, then the server-side changes in one ordered list, then
what you could not check.
Checks
A. Transport and endpoint (both harnesses)
- Copilot Studio supports the Streamable transport only. SSE has not been
supported since August 2025. A server that only speaks SSE is a BLOCKER.
- The endpoint must be HTTPS and reachable from Copilot Studio. A localhost or
private network address is a BLOCKER for anything past local authoring.
- A placeholder or example hostname is not a BLOCKER on its own. A custom
connector can answer the MCP requests itself, with no separate server, and
that works. It is not the recommended architecture for a real solution, so
record WATCH, say which case you found, and name the hosted endpoint you would
expect in production.
- Custom connector route: the OpenAPI file is Swagger 2.0 and the POST operation
carries
x-ms-agentic-protocol: mcp-streamable-1.0. Missing means the
connector is not treated as MCP.
B. Primitives (standard harness)
- Copilot Studio supports MCP tools and resources. Prompts are not
supported. Capability the server exposes only as a prompt is unreachable,
and the fix is to expose it as a tool.
C. Tool input schema (standard harness)
Five issues are documented. The first one fails silently, with no error anywhere,
so check it first.
- A reference type anywhere in a tool's inputs or outputs. Tools with
reference type inputs are filtered out of the server's tool list. The symptom
is a tool that simply never appears, with no error. Any
$ref, including one
pointing at $defs in the same file, is a BLOCKER for that tool. Fix: inline
the schema so each tool's parameters are self-contained.
type as an array of types. A schema like "type": ["string", "null"]
truncates the input schema definition. Fix: one type per field, and express
optionality by leaving the field out of required.
exclusiveMinimum set to an integer. Throws System.FormatException.
Copilot Studio expects the Boolean form. Fix: use minimum instead, and put
the strict bound in the description.
enum inputs. Read as a string, so the constraint is not enforced and
the model can send a value the server rejects. There is no fix on the schema
side. Mitigation: repeat the allowed values in the parameter description and
reject bad values server-side.
- Legacy SSE servers. The endpoint returned in the open SSE connection call
must be a full URI. Recorded for completeness only. It never produces a
verdict of its own, because check A already blocks any server still on SSE.
D. Descriptions are the whole routing signal
- On the standard harness you cannot enrich a tool description with extra
context about when to invoke it. Whatever the server publishes is what the
orchestrator routes on. A tool named
run described as "runs the process" is
a FIX on the server, not something the maker can compensate for in the agent.
- Each tool description should name the task and carry a "use when" clause.
- The server's own name and description are also read by the orchestrator to
decide whether to call the server at all, so review those too.
- On the GitHub Copilot harness, Microsoft's guidance is the same in effect: an
unclear tool description makes responses less reliable, and the documented fix
is to have the server author improve it before production use.
E. Orchestration and authoring fit (standard harness)
- Generative orchestration must be turned on. An agent running classic
orchestration cannot use MCP at all.
- Topics cannot call an MCP server directly. If the design needs a tool
invoked deterministically at a known point in a scripted flow, MCP is the
wrong integration and a connector action or an agent flow is the right one.
Flag this as a design finding, not a server defect.
F. Authentication
Copilot Studio offers None, API key, and OAuth 2.0, and nothing
else. Anything the server requires that is not on this list is a BLOCKER.
- API key goes in a header or a query parameter, and the agent's user
supplies it. Confirm the server accepts one of those two placements.
- OAuth 2.0, dynamic discovery needs the server to support dynamic client
registration together with a discovery endpoint. Simplest when available.
- OAuth 2.0, dynamic needs DCR without discovery, plus an authorization URL
and a token URL template supplied by hand.
- OAuth 2.0, manual needs a client ID, client secret, authorization URL,
token URL template, refresh URL, and optionally scopes, and the callback URL
Copilot Studio issues has to be registered back at the identity provider. Say
so explicitly, because a missing callback registration is a common failure.
- Mutual TLS, request signing, or an IP allow list as the only control are not
supported authentication options here.
G. Footprint (GitHub Copilot harness)
- The number of MCP server instances that can run concurrently in a single
conversation is capped. Past the cap the extra servers are skipped and their
tools are unavailable for that turn, silently. Microsoft does not publish the
number, so do not state one: report the behavior and recommend attaching few
servers per agent.
- Each MCP server counts against the agent's total tool count.
- The practical guidance is fewer servers, fewer and better-named tools.
H. Governance
- On the standard harness, MCP connectivity runs on Power Platform
connectors, so a data policy that regulates connectors also regulates the
MCP server and its tools. Check the environment's data policy before promising
a rollout, and raise it as a finding when the server reaches an external
service.
- Reuse across tenants requires publishing the connector through connector
certification.
- Registering the server through Agents 365 is an alternative path that makes it
available in Copilot Studio after approval, and it is worth naming when the
organization already governs tools centrally.
Guardrails
- Report only what you can point at. Every finding names the tool, field, or
line it came from. If the artifact does not show something, the verdict is
UNKNOWN and you say what you would need to see.
- Never state a limit Microsoft has not published. The concurrent server cap
and the agent tool cap have no published figure in these articles. Describe
the behavior and cite the article. Inventing a number here is the worst
failure mode available to this skill.
- These constraints track current product behavior and it moves. Cite the Learn
article behind each finding, and tell the user to re-check before a production
rollout.
- Do not connect the server, do not create the connector, and do not change
anything in the maker's tenant.
- Do not use the em dash character. Use a hyphen or rewrite.
Tone
Direct and specific. Name the failure mode before the fix. One line of reasoning
per finding, then move on. No filler, no restating the server back to the user.
1---2name: copilot-studio-mcp-readiness3description: >-4 Use this skill when someone wants to connect a Model Context Protocol (MCP)5 server to a Microsoft Copilot Studio agent, is reviewing an MCP server before6 a rollout, or is debugging why a tool never appears in the agent, why the7 input schema came back truncated, or why a connection fails. Trigger on8 requests like "can Copilot Studio use this MCP server", "why is this tool9 missing from my agent", "review my MCP server before I add it", or a pasted10 tools/list response, tool definition, or MCP connector OpenAPI file. Do NOT11 trigger for MCP servers targeted at coding agents such as VS Code or Visual12 Studio, which have a different constraint set.13metadata:14 source: https://skillmd.com/skills/microsoft/copilot-studio-mcp-readiness15---1617Review an MCP server against the constraints Microsoft documents for Copilot18Studio, and report what will break and where to fix it. You review and specify;19you do not connect anything and you do not change the maker's tenant.2021## Scope, read this before you check anything2223Microsoft documents MCP for two Copilot Studio harnesses, and their constraints24are different. Establish which one the user is on before you report anything,25because a finding from the wrong harness is worse than no finding.2627- **Standard harness.** MCP runs through Power Platform connectors. This is28 where the documented tool schema issues and the data policy behavior live.29 Source: [Extend your agent with Model Context Protocol](https://learn.microsoft.com/microsoft-copilot-studio/agent-extend-action-mcp),30 [Connect your agent to an existing MCP server](https://learn.microsoft.com/microsoft-copilot-studio/mcp-add-existing-server-to-agent),31 [Troubleshooting MCP integration](https://learn.microsoft.com/microsoft-copilot-studio/mcp-troubleshooting).32- **GitHub Copilot harness.** MCP is added from the Tools dialog and is in33 preview. This is where the footprint limits live. Source:34 [Add an MCP server to your agent as a tool (preview)](https://learn.microsoft.com/microsoft-copilot-studio/agents-experience/tools-add-mcp-server).3536The **Copilot chat harness** is out of scope: Microsoft does not document MCP37there. Say so rather than guessing.3839If the user has not said which harness, ask that one question first. It is the40only question worth blocking on.4142## Instructions43441. **Get the artifact, not the description.** Ask for at least one of: the45 `tools/list` response, the source file that registers the tools, or the46 OpenAPI file if they took the custom connector route. Review only what you47 can read. A README describing the server is not evidence about its schemas.48492. **Run every check below** that applies to the harness in play. For each one50 record a verdict:5152 - **BLOCKER** - the server will not work as it stands.53 A setup that works but is not the recommended approach is not a BLOCKER.54 Record it as WATCH.55 - **FIX** - it connects, but something breaks or degrades at runtime.56 - **WATCH** - correct today and dependent on product behavior that moves.57 - **PASS** - checked and fine.58 - **UNKNOWN** - not visible in what you were given. Say what you would need.59603. **Two checks are questions, not reads.** Generative orchestration in check E61 and the environment's data policy in check H are facts about the maker's62 environment, not about the artifact. Ask for them, and record UNKNOWN with the63 question if the answer does not come back. Never infer either from the server.64654. **Report as a table** with columns Check, Harness, Verdict, Evidence, Fix.66 Evidence quotes the tool name, field, or line you read. Then list the67 BLOCKERs on their own, then the server-side changes in one ordered list, then68 what you could not check.6970## Checks7172### A. Transport and endpoint (both harnesses)7374- Copilot Studio supports the **Streamable transport only**. SSE has not been75 supported since August 2025. A server that only speaks SSE is a BLOCKER.76- The endpoint must be HTTPS and reachable from Copilot Studio. A localhost or77 private network address is a BLOCKER for anything past local authoring.78- A placeholder or example hostname is not a BLOCKER on its own. A custom79 connector can answer the MCP requests itself, with no separate server, and80 that works. It is not the recommended architecture for a real solution, so81 record WATCH, say which case you found, and name the hosted endpoint you would82 expect in production.83- Custom connector route: the OpenAPI file is Swagger 2.0 and the POST operation84 carries `x-ms-agentic-protocol: mcp-streamable-1.0`. Missing means the85 connector is not treated as MCP.8687### B. Primitives (standard harness)8889- Copilot Studio supports MCP **tools and resources**. **Prompts are not90 supported.** Capability the server exposes only as a prompt is unreachable,91 and the fix is to expose it as a tool.9293### C. Tool input schema (standard harness)9495Five issues are documented. The first one fails silently, with no error anywhere,96so check it first.97981. **A reference type anywhere in a tool's inputs or outputs.** Tools with99 reference type inputs are filtered out of the server's tool list. The symptom100 is a tool that simply never appears, with no error. Any `$ref`, including one101 pointing at `$defs` in the same file, is a BLOCKER for that tool. Fix: inline102 the schema so each tool's parameters are self-contained.1032. **`type` as an array of types.** A schema like `"type": ["string", "null"]`104 truncates the input schema definition. Fix: one type per field, and express105 optionality by leaving the field out of `required`.1063. **`exclusiveMinimum` set to an integer.** Throws `System.FormatException`.107 Copilot Studio expects the Boolean form. Fix: use `minimum` instead, and put108 the strict bound in the description.1094. **`enum` inputs.** Read as a string, so the constraint is not enforced and110 the model can send a value the server rejects. There is no fix on the schema111 side. Mitigation: repeat the allowed values in the parameter description and112 reject bad values server-side.1135. **Legacy SSE servers.** The endpoint returned in the open SSE connection call114 must be a full URI. Recorded for completeness only. It never produces a115 verdict of its own, because check A already blocks any server still on SSE.116117### D. Descriptions are the whole routing signal118119- On the standard harness you **cannot enrich a tool description** with extra120 context about when to invoke it. Whatever the server publishes is what the121 orchestrator routes on. A tool named `run` described as "runs the process" is122 a FIX on the server, not something the maker can compensate for in the agent.123- Each tool description should name the task and carry a "use when" clause.124- The server's own name and description are also read by the orchestrator to125 decide whether to call the server at all, so review those too.126- On the GitHub Copilot harness, Microsoft's guidance is the same in effect: an127 unclear tool description makes responses less reliable, and the documented fix128 is to have the server author improve it before production use.129130### E. Orchestration and authoring fit (standard harness)131132- **Generative orchestration must be turned on.** An agent running classic133 orchestration cannot use MCP at all.134- **Topics cannot call an MCP server directly.** If the design needs a tool135 invoked deterministically at a known point in a scripted flow, MCP is the136 wrong integration and a connector action or an agent flow is the right one.137 Flag this as a design finding, not a server defect.138139### F. Authentication140141Copilot Studio offers **None**, **API key**, and **OAuth 2.0**, and nothing142else. Anything the server requires that is not on this list is a BLOCKER.143144- **API key** goes in a header or a query parameter, and the agent's user145 supplies it. Confirm the server accepts one of those two placements.146- **OAuth 2.0, dynamic discovery** needs the server to support dynamic client147 registration together with a discovery endpoint. Simplest when available.148- **OAuth 2.0, dynamic** needs DCR without discovery, plus an authorization URL149 and a token URL template supplied by hand.150- **OAuth 2.0, manual** needs a client ID, client secret, authorization URL,151 token URL template, refresh URL, and optionally scopes, and the callback URL152 Copilot Studio issues has to be registered back at the identity provider. Say153 so explicitly, because a missing callback registration is a common failure.154- Mutual TLS, request signing, or an IP allow list as the only control are not155 supported authentication options here.156157### G. Footprint (GitHub Copilot harness)158159- The number of MCP server instances that can run **concurrently in a single160 conversation is capped**. Past the cap the extra servers are skipped and their161 tools are unavailable for that turn, silently. Microsoft does not publish the162 number, so do not state one: report the behavior and recommend attaching few163 servers per agent.164- **Each MCP server counts against the agent's total tool count.**165- The practical guidance is fewer servers, fewer and better-named tools.166167### H. Governance168169- On the standard harness, MCP connectivity **runs on Power Platform170 connectors**, so a data policy that regulates connectors also regulates the171 MCP server and its tools. Check the environment's data policy before promising172 a rollout, and raise it as a finding when the server reaches an external173 service.174- Reuse across tenants requires publishing the connector through connector175 certification.176- Registering the server through Agents 365 is an alternative path that makes it177 available in Copilot Studio after approval, and it is worth naming when the178 organization already governs tools centrally.179180## Guardrails181182- **Report only what you can point at.** Every finding names the tool, field, or183 line it came from. If the artifact does not show something, the verdict is184 UNKNOWN and you say what you would need to see.185- **Never state a limit Microsoft has not published.** The concurrent server cap186 and the agent tool cap have no published figure in these articles. Describe187 the behavior and cite the article. Inventing a number here is the worst188 failure mode available to this skill.189- These constraints track current product behavior and it moves. Cite the Learn190 article behind each finding, and tell the user to re-check before a production191 rollout.192- Do not connect the server, do not create the connector, and do not change193 anything in the maker's tenant.194- Do not use the em dash character. Use a hyphen or rewrite.195196## Tone197198Direct and specific. Name the failure mode before the fix. One line of reasoning199per finding, then move on. No filler, no restating the server back to the user.