CtrlK
BlogDocsLog inGet started
Tessl Logo

learn-mcp

Focused interactive tutor for the Model Context Protocol (MCP) path in AI Engineering from Scratch. Start or resume this route when a learner wants to build, secure, debug, verify, or operate MCP clients, servers, transports, gateways, registries, or conformance gates. Teaches one lesson per invocation and records wire evidence in MCP-LEARNING.md.

75

Quality

94%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

88%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.

An exceptionally actionable, well-sequenced instructional skill: exact commands, paths, a state template, explicit gating, and feedback loops for every failure path, with no concept re-teaching. The main costs are the inline 17-row progress template and the lengthy legacy-filename migration procedure, both of which pad token usage and could live in reference files.

Suggestions

Move the blank MCP-LEARNING.md progress template into a reference file (e.g. references/state-template.md) and keep only the route header and a one-line pointer in SKILL.md, trimming ~40 lines from every invocation.

Condense the three-step legacy-filename migration into a short rule ("rename MCP-ENGINEERING-LEARNING.md to MCP-LEARNING.md, verifying content is preserved byte for byte before removing the legacy file") — the atomic-rename fallback detail is more procedure than the tutor needs inline.

Tighten the Lesson 06 walkthrough by referencing the observable fields as a compact list instead of a full paragraph, since the lesson's own manifest checkpointEvidence already enumerates them.

DimensionReasoningScore

Conciseness

The body is dense, imperative, and assumes Claude's competence — it never re-teaches what MCP is. However, the 17-row blank progress table template and the extended legacy-filename migration procedure carry inline bulk that could be trimmed or moved to a reference file, matching anchor 4 ("efficient; minor instances of over-explanation that could be trimmed") rather than 5.

4 / 5

Actionability

Fully executable guidance: exact commands ("python3 phases/13-tools-and-protocols/06-mcp-fundamentals/code/main.py"), exact paths ("learning-paths/model-context-protocol.json", "docs/en.md", "quiz.json"), a raw GitHub fallback URL, a copy-paste state-file template, and a concrete first-lesson script with named observable fields. Copy-paste ready with specific handling for common failure cases.

5 / 5

Workflow Clarity

Clear multi-step sequence (evidence mode → state migration → first lesson → teach → close) with explicit validation and feedback loops throughout: verify-before-remove in the legacy rename, "If the command cannot run... Keep the command checkpoint pending", "If any required evidence is missing, teach or rerun Lesson 15", and Done-gating on checkpoint evidence plus quiz. Destructive file operations are wrapped in validation, so no cap applies.

5 / 5

Progressive Disclosure

No bundle files exist (references/, scripts/, assets/ are all absent), so everything lives in one file. Sections are clearly organized and detail is delegated one level deep to the manifest and repository files rather than duplicated, but the ~45-line blank state-file template is inline bulk that belongs in a separate reference file — matching anchor 4's "minor organization gaps" rather than 5's clean split.

4 / 5

Total

18

/

20

Passed

Description

96%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 voice, explicit use-when clause with concrete trigger verbs, and a comprehensive enumeration of MCP capabilities plus a concrete artifact (MCP-LEARNING.md). The only weakness is mild overlap risk with general MCP build/debug requests from non-learner contexts.

DimensionReasoningScore

Specificity

Enumerates multiple concrete capabilities across the whole MCP surface ("build, secure, debug, verify, or operate MCP clients, servers, transports, gateways, registries, or conformance gates") plus concrete mechanics ("Teaches one lesson per invocation and records wire evidence in MCP-LEARNING.md"). Comprehensive rather than just "several actions", so it matches the 5 anchor over 4.

5 / 5

Completeness

Explicitly answers both: what ("Focused interactive tutor... Teaches one lesson per invocation and records wire evidence") and when ("Start or resume this route when a learner wants to build, secure, debug, verify, or operate...") with concrete trigger phrases, exactly matching the 5 anchor.

5 / 5

Trigger Term Quality

Covers natural user phrasing with both the acronym and full name ("Model Context Protocol (MCP)", "MCP clients, servers") plus verb-level triggers (build, secure, debug, verify, operate). Synonym coverage is broad enough to match the 5 anchor rather than 4's "a few natural terms missing".

5 / 5

Distinctiveness Conflict Risk

The tutor framing tied to a specific curriculum is a clear niche, but the trigger verbs overlap with a generic MCP-development skill — a user asking to "build an MCP server" for production use could pull this in. Mostly distinct with minor overlap risk, matching 4; not 5 because that adjacent-trigger overlap exists.

4 / 5

Total

19

/

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
rohitg00/ai-engineering-from-scratch
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.