CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

use-dedalus-mcp

Use when developers need to build MCP (Model Context Protocol) servers with dedalus_mcp—covers creating tools via @tool, exposing resources via @resource, defining user prompts via @prompt, using context utilities (progress, info, cancellation), and registering components modularly with collect().

Quality

84%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

78%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, example-driven skill body: nearly all code is runnable, the build-test-serve flow is clear, and advanced features are properly split into real reference files. The main gaps are a small amount of already-known MCP-concept explanation, a few snippets with undefined helper functions or missing imports, and the absence of an explicit post-serve verification step.

Suggestions

Remove or shrink the 'Server Primitives' table — MCP's tool/resource/prompt roles are concepts Claude already knows, and the table's token budget is better spent elsewhere.

Make the resource and resource-template snippets fully executable by adding the missing imports (json, Path) and either defining or clearly stubbing helper functions like load_users() and fetch_user().

Add a post-serve verification step (e.g., curl the /mcp endpoint or assert the tool list) so the run workflow has an explicit validation checkpoint.

DimensionReasoningScore

Conciseness

The body is dominated by terse, framework-specific code and one-line explanations ('Decorate, register, serve.', 'Dedalus decouples decoration from registration:'), matching the anchor 'Efficient; minor instances of over-explanation that could be trimmed'. It is not a 5 because the 'Server Primitives' table re-explains what MCP tools, resources, and prompts are — concepts Claude already knows — and lines like 'Tools are just functions' add mild padding. It is well above the score-3 anchor, which expects unnecessary explanation woven into the working guidance.

4 / 5

Actionability

Most guidance is executable and copy-paste ready: the install command, Quick Start (runs end-to-end at http://127.0.0.1:8000/mcp), tools, context access, registration, running, and testing sections are complete. It is not a 5 because several resource/prompt snippets contain undefined helpers ('load_users()', 'fetch_user(user_id)') and missing imports ('json', 'Path'), placing it at the 'concrete code or commands with minor gaps' anchor rather than fully executable coverage.

4 / 5

Workflow Clarity

The build flow is clearly sequenced (Installation → Quick Start → define primitives → register with collect() → serve → test), and the Testing section provides explicit verification ('assert "add" in names'). This matches 'Clear sequence with most checkpoints present; minor validation gaps'. It is not a 5 because there is no explicit post-serve verification or error-recovery guidance (e.g., checking the endpoint responds or what to do when serve fails), and not a 3 because sequence and checkpoints are far more complete than the 'sequence present but checkpoints missing' anchor.

4 / 5

Progressive Disclosure

The body is a clear overview of core primitives with advanced capabilities deferred to well-signaled, one-level-deep references that all exist in the bundle: '**Elicitation**: See references/elicitation.md', '**Sampling**: See references/sampling.md', '**Client**: See references/client.md' — matching the anchor 'Clear overview with well-signaled one-level-deep references; content appropriately split; easy navigation'. The core usage legitimately belongs inline, and no content is buried or nested.

5 / 5

Total

17

/

20

Passed

Description

83%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: third-person, specific, and explicit about both capabilities and trigger conditions. The action list is comprehensive and framework-specific, giving it excellent distinctiveness. Its only weakness is a single, somewhat narrow trigger phrase that could enumerate more natural user phrasings.

DimensionReasoningScore

Specificity

The description lists multiple concrete actions with the exact API surface: 'creating tools via @tool', 'exposing resources via @resource', 'defining user prompts via @prompt', 'using context utilities (progress, info, cancellation)', and 'registering components modularly with collect()' — comprehensive coverage of the framework's capabilities. It clearly matches the anchor 'Lists multiple specific concrete actions; comprehensive coverage' rather than the score-4 anchor, which expects minor gaps; all three MCP primitives plus context and registration are named.

5 / 5

Completeness

Both 'what' ('covers creating tools via @tool, exposing resources via @resource...') and 'when' ('Use when developers need to build MCP... servers with dedalus_mcp') are explicitly present, matching the anchor 'Has both what and when'. It is not a 5 because the when-clause offers a single trigger phrase ('need to build MCP servers') rather than the multiple concrete trigger variations the score-5 anchor exemplifies (e.g., 'or when the user mentions tools, resources, prompts').

4 / 5

Trigger Term Quality

Natural terms a developer would say are present: 'build MCP servers', 'MCP (Model Context Protocol)', 'tools', 'resources', 'prompts', and 'progress' — good keyword coverage matching the anchor 'Good keyword coverage; a few natural terms missing'. It falls short of the score-5 anchor because it lacks broader synonym coverage (e.g., 'expose endpoints', 'agent integrations', 'MCP client tools'), though 'Model Context Protocol' being spelled out partially compensates.

4 / 5

Distinctiveness Conflict Risk

The named framework ('dedalus_mcp'), decorator names ('@tool', '@resource', '@prompt'), and the registration call ('collect()') establish a clear niche with distinct triggers and minimal conflict risk, matching the score-5 anchor. It is not a 4 because no closely related skill would plausibly trigger on these framework-specific terms; the MCP-server domain itself is narrowly scoped.

5 / 5

Total

18

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

adamsardo/dedalus-skills

Path

skills/use-dedalus-mcp/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.