Metadata-Version: 2.4
Name: agentcred
Version: 0.1.0
Summary: Agent Credit Bureau — trust scoring and audit reports for AI agents and MCP servers
Project-URL: Homepage, https://github.com/LuciferForge/agentcred
Project-URL: Repository, https://github.com/LuciferForge/agentcred
Author: LuciferForge
License-Expression: MIT
Keywords: agent,ai,audit,kya,mcp,reputation,security,trust
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Security
Requires-Python: >=3.10
Requires-Dist: kya-agent>=0.2.0
Provides-Extra: full
Requires-Dist: ai-injection-guard>=0.1.0; extra == 'full'
Requires-Dist: mcp-security-audit>=0.2.0; extra == 'full'
Description-Content-Type: text/markdown

# AgentCred

Agent Credit Bureau — trust scoring and audit reports for AI agents and MCP servers.

## Install

```bash
pip install agentcred
```

## Quick Score

```bash
agentcred quick agent-card.kya.json
```

## Full Report

```bash
agentcred report agent-card.kya.json -o report.md
agentcred report agent-card.kya.json --json -o report.json
agentcred report agent-card.kya.json --github-stars 150 --audit-score 85
```

## What It Scores

| Bucket | Weight | Signals |
|--------|--------|---------|
| Identity Completeness | 20% | KYA card, signing, fields, owner, capabilities |
| Security Posture | 25% | Audit history, injection testing, CVEs, compliance |
| Reliability | 35% | Logging, kill switch, rate limits, oversight |
| Behavioral Reputation | 20% | Age, dependencies, GitHub stars, version maturity |

Output: **Score (0-100) + Grade (A-F) + Confidence (minimal/low/medium/high)**

## Part of the LuciferForge Ecosystem

Built on [kya-agent](https://pypi.org/project/kya-agent/) (agent identity standard). Enhanced with [mcp-security-audit](https://pypi.org/project/mcp-security-audit/) and [ai-injection-guard](https://pypi.org/project/ai-injection-guard/).

## License

MIT
