feat: add invocation benchmark package#2366
Open
fahreddinozcan wants to merge 5 commits intomasterfrom
Open
Conversation
TypeScript port of the orchestrator eval script as a proper monorepo package. Benchmarks trigger accuracy across MCP and CLI integration modes with prod/dev comparison support (prod reads from master, dev reads from working tree).
- Add nia:prod and vs:mcp modes for competitive benchmarking - Add versus detection to track which provider Claude picks - Move trigger-eval.json into benchmark package, gitignore results - Remove context7-mcp skill during cleanup to prevent test leaks - Fix nia MCP registration (arg ordering for variadic --header) - Add 15 new eval queries: deep-dive, source-code, research, github-search, niche-lib categories (75 total) - Show provider breakdown in final comparison for versus modes
enesgules
reviewed
Mar 31, 2026
packages/benchmark/src/detection.ts
Outdated
Comment on lines
+16
to
+35
| const NIA_TOOLS = new Set([ | ||
| "search_documentation", | ||
| "search_codebase", | ||
| "index", | ||
| "regex_search", | ||
| "manage_resource", | ||
| "get_github_file_tree", | ||
| "nia_web_search", | ||
| "nia_deep_research_agent", | ||
| "read_source_content", | ||
| ]); | ||
|
|
||
| function isContext7Tool(name: string): boolean { | ||
| return ( | ||
| name.includes("resolve-library-id") || | ||
| name.includes("resolve_library_id") || | ||
| name.includes("query-docs") || | ||
| name.includes("query_docs") | ||
| ); | ||
| } |
Collaborator
There was a problem hiding this comment.
we have to update these lists when there are new tools or tool names change? :'( also i remember nia having more tools, is this list accurate
enesgules
approved these changes
Mar 31, 2026
- Add vs:cli mode: Context7 CLI prod setup + Nia skill side by side - Pull Nia skill from nozomio-labs/nia-skill repo (cached locally) - Detect Nia skill triggers via Skill(nia) and Bash(nia) calls - Fix versus detection to recognize CLI/skill triggers, not just MCP
Use MCP server namespace prefixes (mcp__context7__, mcp__nia__) and naming conventions (nia_*) instead of maintaining a hardcoded set of tool names that breaks when tools are added or renamed.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
packages/benchmark-- a TypeScript port of the Python orchestrator eval script as a proper monorepo packagemaster, dev reads from working tree and uses local MCP buildpnpm benchfrom repo rootModes
mcp:prodmcp:devcli:prodcli:devAd-hoc modes also available:
mcp,mcp+rule,mcp+claude.md,cli+skill,cli+rule,cli+claude.md