CtrlK
BlogDocsLog inGet started
Tessl Logo

talk-firtman-web-mcp-agentic-web

Use when the user asks about Maximiliano Firtman's ("Maxi") AI Native DevCon talk on Web MCP and the agentic web — including questions about how Web MCP differs from MCP, exposing frontend tools to AI agents, the imperative vs declarative Web MCP APIs, tool design guidance (name/description/input schema/execute), the Chrome 149 origin trial, using Chrome DevTools MCP or Puppeteer as a bridge today, the Doom demo, or applying Firtman's "pick one high-value page state" adoption approach to current work.

71

Quality

88%

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

85%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, actionable reference skill with clear sequenced workflows and strong source-grounding guardrails. Its main weakness is that the body repeatedly references outline.md, transcript.md, and quote.md, none of which are present in the bundle, so the progressive-disclosure layer is non-functional.

Suggestions

Ship the referenced bundle files (outline.md, transcript.md, quote.md) under references/ so the progressive-disclosure navigation resolves to real content.

Consolidate the duplicated garbled-word lists (rules 5 and 6) into a single example table to tighten conciseness.

Consider collapsing the repeated "read outline.md → read transcript.md" preamble shared across the five task subsections into one shared grounding step to reduce redundancy.

DimensionReasoningScore

Conciseness

The body assumes Claude's competence (no padding about what MCP/DOM are) and gives a lean framing paragraph, but the read-outline-then-transcript pattern is repeated across five subsections and the garbled-word examples overlap between rules 5 and 6, leaving minor trim opportunities. Not 5 due to this redundancy; not 3 because it is mostly efficient.

4 / 5

Actionability

Concrete, specific guidance throughout: exact file targets, numbered steps, prescribed phrasings ("the talk doesn't address this"), citation by line range, and a precise tool-object spec (name/description/inputSchema/execute) with example quotes. As an instruction-only skill the guidance is fully actionable and covers the common cases.

5 / 5

Workflow Clarity

Each task type has a clear numbered sequence with explicit verification checkpoints against the source (locate in outline → read transcript range → answer with quotes → cite lines → if absent, say so) and guardrail feedback ("If the framework genuinely doesn't fit, say so"; "Do not over-cite... stay quiet"). No destructive/batch ops, so the 3-cap does not apply.

5 / 5

Progressive Disclosure

The design is sound — one-level-deep, clearly signaled references (outline.md, transcript.md, quote.md) with content correctly externalized — but those referenced files are absent from the bundle, so the overview points to non-existent detailed materials and the navigation does not function. Not 4 because broken reference targets are more than a minor organization gap; not 2 because the structure and signaling themselves are good.

3 / 5

Total

17

/

20

Passed

Description

91%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, highly specific description with comprehensive trigger keywords and minimal conflict risk. Its only weakness is that the "what does this skill do" action is implied rather than explicitly led with an action verb.

Suggestions

Lead with an explicit action verb for the "what" (e.g., "Answers questions about and applies Maximiliano Firtman's ... talk on Web MCP ...") before the "Use when..." clause so both what and when are explicitly stated.

DimensionReasoningScore

Specificity

Lists multiple specific concrete subtopics — "how Web MCP differs from MCP", "imperative vs declarative Web MCP APIs", "tool design guidance (name/description/input schema/execute)", "Chrome 149 origin trial", "the Doom demo" — giving comprehensive coverage of the talk's scope rather than vague language.

5 / 5

Completeness

The "when" is exemplary and concrete ("Use when the user asks about..." with detailed triggers), but the "what" action is only implied through "asks about" rather than explicitly stated with a verb like "Explains/Answers". Not 5 because the anchor requires both what AND when to be clearly and explicitly stated; not 3 because the trigger guidance is far from weak.

4 / 5

Trigger Term Quality

Covers the natural terms a user would say — "Web MCP", "MCP", "AI agents", "Chrome DevTools MCP", "Puppeteer", "agentic web", "Doom demo", plus the speaker and conference names — with strong synonym/variant coverage. Not a 4 because the keyword set is comprehensive rather than just 'good with a few missing'.

5 / 5

Distinctiveness Conflict Risk

Pins a clear niche — a named speaker (Firtman/"Maxi"), named conference (AI Native DevCon), and a specific talk topic — so triggers are distinct and conflict risk with other skills is minimal.

5 / 5

Total

19

/

20

Passed

Validation

93%

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

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

Total

15

/

16

Passed

Repository
jscraik/Agent-Skills
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.