# NationGraph Claude Code API Evaluation

**Category:** Public Sector Intelligence
**Score:** 22 (D)

This is the agent-readable Devtool Arena evaluation page for NationGraph on Claude Code API.

## Summary

| Metric | Value |
|--------|-------|
| Status | completed |
| Overall score | 22 |
| Grade | D |
| Eval score | 26 |
| Discovery score | 14 |
| Cost | $1.00 |
| Runtime | 8m 44s |
| Tool calls | 42 |
| Errors | 6 |
| Tokens used | 1789212 |
| Success | Yes |

## Agent-Readiness Checklist

| Signal | Evidence |
|--------|----------|
| Context7 | Not found |
| llms.txt | https://nationgraph.com/llms.txt |
| MCP server | Not found |
| Typed SDK | Not found |
| OpenAPI | Not found |
| Agent skills | Not found |
| CLI | Not found |

## Evaluation Prompt

````text

# Task
Using https://api.nationgraph.com/docs, complete the following task:

Using NationGraph, build me a script that runs a small public-sector procurement research workflow:

1. Run TWO related lookups against NationGraph:
   a. A broad query for public-sector purchasing activity in "cybersecurity" (contracts, bids, spending, or awards — whatever NationGraph exposes)
   b. A more specific, recency-focused query for "cybersecurity" awards or contracts in 2024-2025
2. Merge the results from both lookups, deduplicate by a stable identifier (award/contract/document ID, or URL if no ID is available), and select the top 5 most relevant records
3. If NationGraph supports fetching/extracting a record's full detail, retrieve the single most relevant record and write a 2-3 sentence summary grounded in that fetched detail (not just the list snippet)
4. Print the output as JSON with two keys:
   - "results": the top 5 deduplicated records, each with a title/name, an agency or buyer (when available), and a url or record id
   - "summary": your grounded 2-3 sentence summary of the top record (include the source url or record id) 

## Execution
After creating the script, run it to verify it works:
```bash
cd /home/daytona/app && python <your_script>.py
```

The script should print output to stdout. If there are errors, debug and fix them until it runs successfully.

````

## Grader Results

| Check | Passed | Weight | Score | Details |
|-------|--------|--------|-------|---------|
| Cost | No | 0 | 0 | $1.00 (≥$1.00, too expensive) |
| Time | No | 0 | 0 | 524s (5–10min, very slow) |
| Efficiency | No | 0 | 0 | 42 tool calls (>25, inefficient) |
| Found Docs | Yes | 0 | — | Touched docs at api.nationgraph.com: yes (27 matching tool calls) |
| Zero Errors | No | 0 | — | Tool outputs with errors/tracebacks: 6 |
| Syntax Valid | Yes | 0 | — | Code compiles/parses correctly |
| Created Files | Yes | 0 | — | Generated files: 1 (need ≥1) |
| Full Execution | No | 0 | — | Method: syntax_check, Exit: 0 |

## Run Artifacts

| Artifact | Value |
|----------|-------|
| Conversation turns | 43 |
| Tool call traces | 42 |
| Generated files | 1 |
| Exit code | 0 |
| Completed at | 2026-09-26T17:55:27.444869+00:00 |

### Generated Files

- /home/daytona/app/nationgraph_procurement_research.py

## Related Pages

- [Claude Code API leaderboard](/leaderboard/claudecode/api)
- [Compare Claude Code API companies](/leaderboard/claudecode/api/compare)
- [Agent Landscape](/leaderboard/discoverability)

Canonical URL: https://devtoolarena.com/claudecode/api/nationgraph