{
  "markdown": "# AEGIS Governance — MCP Server\n\n[![PyPI](https://img.shields.io/pypi/v/aegis-governance)](https://pypi.org/project/aegis-governance/)\n[![License: BSL-1.1](https://img.shields.io/badge/license-BSL--1.1-blue)](https://aegis.undercurrentholdings.com/docs)\n\n**Quantitative governance for AI agents and engineering decisions.** AEGIS evaluates proposals through six quantitative gates — Risk, Profit, Novelty, Complexity, Quality, Utility — and returns a structured decision (`PROCEED` / `PAUSE` / `HALT` / `ESCALATE`) with confidence scores, rationale, and a hash-chained audit trail.\n\nGive your agent a decision gate it can call before it acts — and an audit record compliance can actually read (NIST AI RMF, EU AI Act Annex IV).\n\n- **Works immediately, no signup**: the local server runs in sandbox mode (10 evaluations/day).\n- 6 local tools (evaluations, risk checks, health, decision history, usage) — 10 on the hosted server.\n- Hosted server with hash-chained audit trails — [free Community tier](https://portal.undercurrentholdings.com/signup?product=aegis&utm_source=github&utm_medium=aegis-mcp) (100 evaluations/month, no credit card).\n- Want to see it before connecting? [Try the Advisor in your browser](https://aegis.undercurrentholdings.com/advisor) — no install, no signup.\n\n## Quickstart (local, no account needed)\n\n```bash\npip install \"aegis-governance[mcp]\"\n```\n\n**Claude Code**\n\n```bash\nclaude mcp add aegis -- aegis-mcp-server\n```\n\n**Cursor** (`.cursor/mcp.json`) / **Windsurf** / any stdio MCP client:\n\n```json\n{\n  \"mcpServers\": {\n    \"aegis\": { \"command\": \"aegis-mcp-server\" }\n  }\n}\n```\n\n**VS Code** (`.vscode/mcp.json`):\n\n```json\n{\n  \"servers\": {\n    \"aegis\": { \"type\": \"stdio\", \"command\": \"aegis-mcp-server\" }\n  }\n}\n```\n\nRuns in sandbox mode out of the box. Set `AEGIS_API_KEY` in the server's\nenvironment ([free key](https://portal.undercurrentholdings.com/signup?product=aegis&utm_source=github&utm_medium=aegis-mcp))\nto unlock decision history, usage reports, and risk checks. Requires Python >= 3.10.\n\n## Hosted server (streamable-http, full 10-tool surface)\n\nGet a free API key at [portal.undercurrentholdings.com](https://portal.undercurrentholdings.com/signup?product=aegis&utm_source=github&utm_medium=aegis-mcp) (GitHub/Google sign-in, key provisioned automatically), then:\n\n**Claude Code**\n\n```bash\nclaude mcp add --transport streamable-http aegis https://mcp.aegis.undercurrentholdings.com/mcp \\\n  --header \"Authorization: Bearer YOUR_API_KEY\"\n```\n\n**Cursor** (`.cursor/mcp.json`) / **Windsurf** / any streamable-http MCP client:\n\n```json\n{\n  \"mcpServers\": {\n    \"aegis\": {\n      \"type\": \"streamable-http\",\n      \"url\": \"https://mcp.aegis.undercurrentholdings.com/mcp\",\n      \"headers\": {\n        \"Authorization\": \"Bearer YOUR_API_KEY\"\n      }\n    }\n  }\n}\n```\n\n**VS Code** (`.vscode/mcp.json`):\n\n```json\n{\n  \"servers\": {\n    \"aegis\": {\n      \"type\": \"http\",\n      \"url\": \"https://mcp.aegis.undercurrentholdings.com/mcp\",\n      \"headers\": {\n        \"Authorization\": \"Bearer YOUR_API_KEY\"\n      }\n    }\n  }\n}\n```\n\n## Prefer a local SDK instead of MCP?\n\nThe Python SDK has a sandbox mode that works with no account at all (10 evaluations/day):\n\n```bash\npip install aegis-governance\n```\n\n```python\nfrom aegis import Aegis\n\ndecision = Aegis().evaluate(\n    proposal_summary=\"Add Redis caching layer to reduce API latency\",\n    risk_baseline=0.02, risk_proposed=0.05,\n    novelty_score=0.75, complexity_score=0.8, quality_score=0.9,\n)\nprint(decision.status)  # \"proceed\"\n```\n\n> The local stdio MCP server above ships in `aegis-governance` >= 1.3.0 via the `[mcp]` extra.\n\n## Tools\n\n| Tool | What it does |\n|------|--------------|\n| `aegis_evaluate_proposal` | Full six-gate evaluation of a proposal; returns PROCEED/PAUSE/HALT/ESCALATE with per-gate scores and rationale |\n| `aegis_quick_risk_check` | Fast risk screen for a proposed change |\n| `aegis_check_thresholds` | Current gate threshold configuration |\n| `aegis_get_scoring_guide` | Domain-specific guidance for deriving gate parameters (e.g. `cicd`) |\n| `aegis_record_proposal` | Record a proposal for later verification |\n| `aegis_list_proposals` | List recorded proposals |\n| `aegis_verify_proposals` | Verify recorded proposals against outcomes |\n| `aegis_list_decisions` | List past governance decisions |\n| `aegis_get_decision` | Fetch a specific decision with full audit detail |\n| `aegis_crypto_status` | Hash-chain audit integrity status |\n\n## Why a governance gate?\n\nAI agents make thousands of decisions with no record of why. AEGIS gives every consequential action a quantitative evaluation and a tamper-evident audit entry — so \"the agent decided to deploy\" becomes a signed, replayable record with gate scores and rationale.\n\n- **Six gates**: Risk, Profit, Novelty, Complexity, Quality, Utility — calibrated thresholds, KL-divergence drift detection\n- **Audit-ready**: hash-chained decision log; NIST AI RMF and EU AI Act Annex IV artifact generation\n- **Five integration surfaces**: MCP (this repo), Python SDK, REST API, CLI, GitHub Action\n\n## Links\n\n- **Docs**: [aegis.undercurrentholdings.com/docs](https://aegis.undercurrentholdings.com/docs) · [MCP tools reference](https://aegis.undercurrentholdings.com/docs/api/mcp-tools)\n- **Try it in the browser** (no install): [AEGIS Advisor](https://aegis.undercurrentholdings.com/advisor)\n- **Pricing**: [portal.undercurrentholdings.com/pricing](https://portal.undercurrentholdings.com/pricing) — free Community tier; paid tiers for teams and regulated environments\n- **Source distribution**: [PyPI `aegis-governance`](https://pypi.org/project/aegis-governance/) (BSL-1.1)\n\n---\n\nBuilt by [Undercurrent](https://undercurrentholdings.com) — *Agency over agents.*\n",
  "bytes": 5705,
  "sha": "9dbce59053e0f609f1caa14d4dccfe515d34aebbf363073c855e6359b8dce564",
  "repo_slug": "undercurrentai/aegis-mcp",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_com_undercurrentholdings_aegis_df4d72c5/readme"
}