summarize

Fetch a URL or convert a local file (PDF/DOCX/HTML/etc.) into Markdown using `uvx markitdown`, optionally it can summarize

Quality

75%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/summarize/SKILL.md

Quality

Content

71%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.

The body is concise, well-sectioned, and highly actionable with exact commands for both conversion and summarization. Its weaknesses are a dangling reference to the `to-markdown.mjs` wrapper (absent from the skill folder), duplicated explanation of the temp-file/"Hint" behavior, and mild redundancy across the three command variants.

Suggestions

Bundle `to-markdown.mjs` in the skill (e.g., `scripts/to-markdown.mjs`) and reference it by explicit path, or drop the wrapper and inline the direct `uvx` + `pi` commands so every command is self-contained.

State the temp-file + "Hint" behavior once (either the Tip or the numbered "This will" list) instead of twice.

Clarify when to use the direct `uvx` command vs. the wrapper (e.g., "use the wrapper when you need the temp file path or a summary") to remove ambiguity between the three command variants.

Dimension	Reasoning	Score
Conciseness	The body is lean and assumes Claude's competence ("Turn 'things' ... into Markdown"), with no concept explanations. Not 5: the temp-file/"Hint" behavior is explained twice (the Tip paragraph and the numbered "This will" list), and three command variants add mild redundancy.	4 / 5
Actionability	Gives exact, executable commands (`uvx --from 'markitdown[pdf]' markitdown <url-or-path>`, the `--summary --prompt` invocation) covering the common cases. Not 5: two of the three primary commands depend on `node to-markdown.mjs`, which is not present in the skill's bundle (no scripts/ directory exists), leaving a concrete gap.	4 / 5
Workflow Clarity	Both flows (convert, convert+summarize) are sequenced clearly, including the explicit "cd here first" instruction and the 3-step summarize pipeline; operations are non-destructive so no validation cap applies. Not 5: the reliance on the unbundled wrapper and the split between direct-uvx and wrapper usage leaves minor ambiguity about which path to use.	4 / 5
Progressive Disclosure	Sections (When to use / Quick usage / Convert / Convert + summarize) are well organized and the body is appropriately short, but the body references `to-markdown.mjs` as "the wrapper below" and that file is not in the bundle listing — a dangling reference. Not 4: a referenced bundle file that does not exist undermines navigation and structure.	3 / 5
	Total	15 / 20 Passed

Description

66%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.

The description is concise, third-person, and specific about what the skill does, with a named tool and concrete formats. Its main weaknesses are the missing "when to use" trigger clause (capping completeness) and limited synonym coverage for trigger terms.

Suggestions

Append an explicit trigger clause, e.g., "Use when the user shares a URL or document (PDF, Word, PowerPoint) to convert, or asks to summarize a web page or file."

Add natural synonyms and extensions to broaden trigger matching: "Word docs (.docx)", "PowerPoints (.pptx)", "webpages", "convert to Markdown".

Fix the awkward trailing clause "optionally it can summarize" — e.g., "...into Markdown using `uvx markitdown`, with optional summarization of the result."

Dimension	Reasoning	Score
Specificity	"Fetch a URL or convert a local file (PDF/DOCX/HTML/etc.) into Markdown using `uvx markitdown`, optionally it can summarize" lists several concrete actions (fetch URL, convert multiple formats, summarize) with a named tool. Not 5: coverage is narrow — a single tool path with no complementary actions (e.g., writing output, processing results).	4 / 5
Completeness	The "what" is clear (fetch/convert to Markdown, optional summarize) but there is no "Use when..." or equivalent explicit trigger clause, which caps completeness at 3 per the judging guideline. Not 4: the "when" is entirely absent rather than merely weakly stated.	3 / 5
Trigger Term Quality	Includes natural user terms — "URL", "PDF", "DOCX", "HTML", "Markdown", "summarize" — matching what a user would say when needing this skill. Not 5: misses common synonyms and extensions like "Word", "webpage", "PowerPoint", ".pdf".	4 / 5
Distinctiveness Conflict Risk	The markitdown/Markdown-conversion framing is a fairly distinct niche with specific triggers (URL fetch, doc-to-Markdown conversion). Not 5: "summarize" and generic file conversion could overlap with other document or summarization skills.	4 / 5
	Total	15 / 20 Passed

Validation

100%

Warnings & errors only

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: mitsuhiko/agent-stuff
Path: skills/summarize/SKILL.md
Commit: 0865c84

Reviewed: 8 days ago

Table of Contents

Discovery Implementation Validation

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.