Public leaderboard

Public assessment

jawdat6/fixgraph-mcp (fixgraph-mcp)

fixgraph-mcp · v1.0.3 · scanned

What changed in the harness

Selection accuracy 100→100, token cost up 6%, unconfirmed writes 0%→0%.

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

0.0 / 30

0.0 out of 30
02Legibility

28.8 / 30

28.8 out of 30
03Economics

20.0 / 20

20.0 out of 20
04Discoverability

12.3 / 20

12.3 out of 20

Highest-impact fix

Estimated gain +30 points

Add explicit identity and permission preflight tools

Expose machine-readable principal/tenant confirmation and a non-mutating permission check so agents can verify both before destructive actions.

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
fixgraph_get_fixes
no_return_description
Get all verified fixes for a specific FixGraph issue by its ID or slug. Returns an array of fix objects (title, root cause, ordered steps, validation notes, and risk level) associated with the issue.
fixgraph_submit_fix
name_restates_behavior no_return_description
Create a new fix entry attached to an existing FixGraph issue, including title, root cause, ordered remediation steps, optional validation instructions, and risk level. Requires the FIXGRAPH_API_KEY env var to be set. Returns the created fix record, including its assigned ID and submission status.

Selection evidence

Confusable tool pairs.

3 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.

Tool A Tool B Confidence Why they collide
fixgraph_search fixgraph_get_fixes medium A task like 'find fixes for this database timeout error' could plausibly be routed to either tool since fixgraph_search returns issues with fix counts while fixgraph_get_fixes returns actual fixes for a known issue; an agent lacking an issue_id might mistakenly call get_fixes with a query string instead of first searching.
fixgraph_search fixgraph_submit_fix low Both relate to fixes but have clearly distinct verbs (search vs submit) and different required inputs (query vs issue_id/title/root_cause/steps), making confusion unlikely for a natural-language task.
fixgraph_get_fixes fixgraph_submit_fix low get_fixes retrieves existing fixes while submit_fix creates a new one; a task like 'add this fix to the issue' vs 'show me the fixes' are distinguishable, so confusion is unlikely though both operate on the same issue_id.

Compare the field

One score is useful.
The evidence makes it actionable.

Back to the leaderboard