Skip to main content
Glama
deenesjakoruzh
alessandro-martelli

content-knowledge-mcp

by alessandro-martelli
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

content-knowledge-mcp

An MCP server that gives a content agent access to a company's editorial knowledge: brand rules, a visual template catalog, and past posts with their outcomes. The knowledge lives in plain markdown files the company owns; the server only exposes it.

Built as the technical core of my BA thesis on AI agents and workflow automation. The version here is the generic engine, with an invented company as example knowledge.

Why there is no vector database

The obvious move is to embed the knowledge and retrieve by similarity. This server deliberately does not, for two reasons.

At this scale, structured search over curated files wins. A company's editorial rules are tens of blocks, not tens of thousands. Splitting them into chunks and retrieving by cosine similarity adds an indexing step, a store to keep in sync and a class of failure that is hard to see, in exchange for recall you already had.

Every rule carries an ID, and that is the point. Rules are R1, R17, R905; templates are T901, T905; launch records are L2. The agent is instructed to cite the IDs it used. That turns each proposal into something a human can audit in seconds: the reviewer reads rules_used: ["R920", "R903"] and knows exactly which rules produced that post. An embedding returns a paragraph; it does not return an argument. When the goal is a human gate that works, being auditable beats being clever.

The cost is real and worth stating: search is lexical, so a query that shares no words with a rule will not find it. The server says NO_MATCH instead of pretending, which is the behaviour the evaluation cases check.

Related MCP server: Publisher Content MCP Server

Architecture

  Claude Desktop / Claude Code ──stdio──┐
                                        ├──▶ content-knowledge-mcp ──▶ KNOWLEDGE_PATH/
  n8n MCP Client Tool ──────────HTTP────┘         (3 tools)              guidelines/*.md   R# rules, L# launch records
       (host.docker.internal:8765)                                       templates/catalog.md   T# catalog
                                                                         past-posts/*.md        past posts and outcomes

Three tools, one transport switch, no state:

Tool

Returns

search_guidelines(query)

matching rule blocks with their R# and L# IDs

list_visual_templates(query)

the template catalog, filtered, with T# IDs and usage rules

get_post_examples(query)

up to three past posts with their outcomes, for tone calibration

The server owns no content. Point KNOWLEDGE_PATH at a different folder and it serves a different company: that is what made it publishable without touching the code.

Quickstart

git clone https://github.com/alessandro-martelli/content-knowledge-mcp
cd content-knowledge-mcp
python3 -m venv .venv && .venv/bin/pip install fastmcp

# run the twelve checks against the example knowledge
KNOWLEDGE_PATH=examples/knowledge .venv/bin/python smoke_test.py

# stdio, for Claude Desktop or Claude Code
KNOWLEDGE_PATH=examples/knowledge .venv/bin/python server.py

# HTTP on :8765, for the n8n MCP Client Tool
KNOWLEDGE_PATH=examples/knowledge .venv/bin/python server.py --http

The example knowledge under examples/knowledge/ is entirely invented: a fictional design school, with rules, launch records, a template catalog and three past posts. It exists so the repo runs for anyone, and so the tests have something to bite on.

Evaluation

Smoke tests: 12/12. Each check goes through a real fastmcp.Client over the protocol, not through the helper functions, so a break in tool registration or transport shows up here.

Agent-level protocol: 8 cases, 5 of them adversarial (conflicting rules, out-of-scope request, tool unavailable, no matching template, underspecified input). Executed August 2026 against the two agents' prompts: 6 of 8 passed, every output was valid JSON on the first try, and the two failures were defects in the test suite rather than in the agent. Both are written up in eval/test-log.md with the fix each one needs. The log is kept as it came out: a suite that only ever reports success is not measuring anything.

Honest limits

  • Search is lexical and scores by term frequency. It does not understand synonyms; a rule phrased in words the query does not use will not be found.

  • The knowledge folder layout is fixed (guidelines/, templates/catalog.md, past-posts/), and the catalog filename is hardcoded.

  • There is no write path and no cache: every call reads the files. That is fine at this size and would not be at ten thousand blocks.

  • The two open defects from the evaluation are not fixed in this code yet: rules_used cannot cite L# records, and the plan schema cannot express a row whose channel is unknown.

Licence

MIT, see LICENSE.

Available Tools

3 tools
get_post_examplesA

Retrieve past posts with their outcomes, for tone calibration. Query by topic, channel or type, e.g. 'linkedin annuncio'. Returns up to 3 examples; never copy sentences from them.

ParametersJSON Schema
NameRequiredDescriptionDefault
queryNo

Output Schema

ParametersJSON Schema
NameRequiredDescription
resultYes

TDQS

A4.5/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description bears the full behavioral disclosure burden. It reveals the return limit ('Returns up to 3 examples') and adds a usage constraint ('never copy sentences from them'). The verb 'Retrieve' sufficiently implies a read-only operation for this kind of tool. It could optionally disclose more about the output shape, but that is already structurally represented by an output schema.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences contain the entire message with no filler. The tool's value is front-loaded, followed by a practical example, a return limit, and a usage constraint. Every clause earns its place.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a single-parameter retrieval tool with an output schema available, the description covers what is returned, how to query, and a practical guardrail. The only omission is a note about the default parameter, but the schema already declares a default. The agent can select and invoke the tool with confidence.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema has 0% description coverage on the only parameter, `query`, so the description must compensate. It does: the parameter can be used to search by topic, channel, or type, and an example value is given. It does not describe what empty input does, though schema's default value covers that reasonable.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description starts with a specific verb and resource: 'Retrieve past posts with their outcomes'. It also adds the intended use, 'tone calibration', which makes the tool's function concrete. It is clearly distinguishable from siblings like search_guidelines and list_visual_templates because neither of those fetches example posts.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description tells the agent that querying can be by topic, channel, or type, with a concrete example ('linkedin annuncio'). It does not explicitly contrast with search_guidelines or list_visual_templates, but the purpose of tone calibration and the example query provide enough context. A score of 5 would require explicit when-to-use versus the sibling tools.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

list_visual_templatesA

List the official visual template catalog with usage rules (T# IDs). Optional query filters by content type, e.g. 'annuncio corso'. Empty query returns the full catalog.

ParametersJSON Schema
NameRequiredDescriptionDefault
queryNo

Output Schema

ParametersJSON Schema
NameRequiredDescription
resultYes

TDQS

A3.7/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations providing read-only or side-effect hints, the description carries the behavioral burden. It communicates that the tool lists an official catalog, returns usage rules with T# IDs, and behaves differently based on the query parameter. It does not go further to describe matching behavior, ordering, or whether any 'rules' are embedded in the returned data, but the core read-oriented behavior is reasonably clear.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is two concise sentences, front-loads the main purpose, and uses an example to clarify parameter usage. There is no filler or repetition of schema fields, and every phrase adds value.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a simple list-with-optional-filter tool, the description covers the tool's purpose, the optional query, and the empty-query behavior. The presence of an output schema compensates for not describing the return structure. Minor ambiguity remains about what 'content type' means and how T# IDs are used, but as a whole it is largely sufficient.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0%, so the description must compensate, and it does. The 'query' parameter is given semantics: it optionally filters by content type, with the example 'annuncio corso', and empty returns the full catalog. This is meaningful beyond the raw string schema. It could be slightly stronger, but it sufficiently explains the one parameter.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states a specific action, 'List the official visual template catalog', with a resource and an explicit mention of what makes these templates usable: the T# IDs. It does not explicitly contrast itself with the sibling tools search_guidelines or get_post_examples, so it stops short of full differentiation, but the verb and resource are unambiguous.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It explains when the query parameter is useful and that an empty query returns the full catalog, which provides practical guidance for using the tool. However, it does not explicitly state when to prefer this tool over sibling tools such as search_guidelines or get_post_examples, so the usage context remains partially implied.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

search_guidelinesB

Search the company brand and editorial rules. Returns matching rule blocks with their IDs (R#), e.g. 'tone course post linkedin'.

ParametersJSON Schema
NameRequiredDescriptionDefault
queryYes

Output Schema

ParametersJSON Schema
NameRequiredDescription
resultYes

TDQS

B3.4/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the burden of indicating behavior safely. It does disclose that the tool returns 'matching rule blocks with their IDs (R#)', which explains the shape of the output. However it omits details such as whether it is read-only, how many results are returned, or any limitations on query syntax.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences only: the first states purpose, the second gives return format and a clear example. Every sentence serves a direct purpose with no fluff. It is appropriately front-loaded with the core action.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a simple one-parameter search tool with an output schema already defined, the description is nearly complete. It explains the input style and output format for typical searches. It still lacks usage guidance relative to siblings, which slightly reduces completeness.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Parameter schema coverage is 0% since the query string has no description. The description compensates by giving a concrete example ('tone course post linkedin'), which clarifies that the query is free-form text and likely case-insensitive phrase style. This is meaningful added meaning beyond the bare schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states a specific action ('Search') and a precise resource ('company brand and editorial rules'). It distinguishes this as a search tool from the sibling tools (list_visual_templates, get_post_examples), though it could be more explicit about exactly how it differs.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description does not say when to use this tool rather than the siblings. It gives no context, precedence, or exclusion. The example query implies usage, but an agent is given no guidance to choose this over the alternatives.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 3 tool updatesv0.1.0
    • First observedget_post_examples
    • First observedlist_visual_templates
    • First observedsearch_guidelines
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A4/5.0

Scored across 3 tools

Disambiguation5/5

Each tool targets a distinct knowledge resource: editorial rules, visual templates, and past post examples. The descriptions clearly distinguish search, list, and get operations despite all supporting query. An agent can reliably choose the correct tool for each content creation step.

Naming Consistency5/5

All tool names follow the consistent snake_case verb_noun pattern: search_guidelines, list_visual_templates, get_post_examples. The verbs also accurately reflect the operation type, making the naming predictable and scannable.

Tool Count5/5

With only three tools, the server is tightly focused on content knowledge retrieval. Each tool covers a core knowledge domain for content creation—rules, templates, and examples—without redundant or unnecessary additions.

Completeness4/5

The set covers the main content-making inputs: brand/editorial guidelines, visual template catalog, and performance-calibrated post examples. Minor gaps exist such as fetching a specific rule by ID or retrieving more than three examples, but the tool set supports its purpose effectively.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • mcp-guide

    deeprave
    A
    license
    B
    quality
    A
    maintenance
    Enables AI agents to access and manage project guidelines, documentation, and context through a structured content system with template support and workflow management.
    37
    MIT

  • Publisher Content MCP Server

    ravintambimuttu-eng
    F
    license
    Not graded
    quality
    D
    maintenance
    Exposes publisher journalism to LLMs and agent frameworks via search, retrieval, source grounding, and consistent citation, enabling accurate content retrieval with attribution and policy enforcement.
    -