Public leaderboard

Public assessment

ofershap/mcp-server-markdown (mcp-server-markdown)

mcp-server-markdown · v1.0.0 · scanned

What changed in the harness

Selection accuracy 100→100, token cost up 4%, unconfirmed writes 100%→100%.

Category breakdown

Where the score comes from.

Earned points across the four signals Gradable measures. Safety and Legibility are scored out of 30; Economics and Discoverability are scored out of 20.

01Safety

30.0 / 30

30.0 out of 30
02Legibility

28.7 / 30

28.7 out of 30
03Economics

20.0 / 20

20.0 out of 20
04Discoverability

17.1 / 20

17.1 out of 20

Highest-impact fix

Estimated gain +3 points

Make target tools discoverable on the first call

Clarify tool names, decision boundaries, and required argument schemas so an agent can choose and construct the target call without exploratory steps.

Description evidence

Defects and rewrites.

2 defects found across the exposed tool descriptions. Suggested rewrites make purpose, inputs, boundaries, and returns easier for an agent to understand.

Tool Defect types Suggested rewrite
list_headings
name_restates_behavior no_return_description
Return a table of contents for a markdown file by listing every heading level (# through ######) in document order with each heading's text and level, so callers can navigate long documents or build an index.
find_code_blocks
name_restates_behavior no_return_description
Find all fenced code blocks in a markdown file and return each block's source text and language tag when present. If a language is provided, only blocks matching that language (e.g. typescript, python) are returned, useful for locating examples of a specific code type.

Selection evidence

Confusable tool pairs.

1 pair where similar names or overlapping descriptions may send an agent toward the wrong tool.

Tool A Tool B Confidence Why they collide
list_files search_docs medium A task like 'find which markdown docs mention installation' or 'which files cover topic X' is plausibly misread as listing files vs full-text searching content, since both take a directory and both descriptions center on '.md files'. The word 'find' in the task points to search_docs, but an agent could pick list_files to enumerate paths instead of content matches.

Compare the field

One score is useful.
The evidence makes it actionable.

Back to the leaderboard