AgentCop API

Submit agent source code for a security scan. Returns a trust score, risk level, and a detailed list of detected issues. Results are cached by code hash — identical code submitted again will return the cached result instantly.

Request body

{
  "code": "import os\nagent = ...",
  "description": "LangChain ReAct agent with web search",
  "agent_name": "my-agent",
  "framework": "langchain"
}

Field	Type	Required	Description
code	string	Yes	Agent source code to scan (max 100 KB)
description	string	No	Human-readable description of the agent
agent_name	string	No	Identifier used in scan history and share URLs
framework	string	No	Hint for framework-specific rules: langchain, crewai, autogen, moltbook

Response

{
  "scan_id": "scan_abc123",
  "trust_score": 72,
  "risk_level": "MODERATE",
  "total_issues": 3,
  "issues": [
    {
      "id": "issue_001",
      "type": "LLM01_PROMPT_INJECTION",
      "severity": "HIGH",
      "description": "f-string interpolation in LLM prompt",
      "line": 14,
      "code_snippet": "prompt = f'Answer: {user_input}'",
      "fix": "Use PromptTemplate with input_variables",
      "owasp": "LLM01",
      "cwe": "CWE-20"
    }
  ],
  "scan_time_ms": 234,
  "created_at": "2026-04-06T14:23:11Z",
  "share_url": "https://agentcop.live/scan/scan_abc123"
}

Field	Type	Description
trust_score	integer (0–100)	Higher is safer. Penalized by severity and count of issues.
risk_level	string	SAFE, LOW, MODERATE, HIGH, CRITICAL
issues[].owasp	string	OWASP LLM Top 10 category (e.g., LLM01)
issues[].cwe	string	CWE identifier
share_url	string	Public permalink to this scan result

Scan an entire agent project by uploading a ZIP archive. Accepts multipart/form-data. All .py files inside the archive are scanned and their issues are merged into a single result. Maximum archive size is 10 MB.

Request (multipart/form-data)

curl -X POST https://api.agentcop.live/api/scan/zip \
  -H "X-AgentCop-Key: your_api_key_here" \
  -F "file=@agents.zip" \
  -F "description=My agent project"

Field	Type	Required	Description
file	file	Yes	ZIP archive containing agent source files
description	string	No	Project description, stored with scan result
framework	string	No	Framework hint applied to all files

The response schema is identical to POST /api/scan, with an additional files_scanned integer field.

Retrieve a previously created scan result by its ID. Scan results are retained for 90 days on Free plans and indefinitely on Pro and Enterprise plans.

Path parameters

Parameter	Type	Description
scan_id	string	The scan ID returned by POST /api/scan

Response schema is identical to POST /api/scan.

Returns an SVG trust badge for a scan result. Intended for embedding in READMEs and documentation. The response carries Content-Type: image/svg+xml and Cache-Control: max-age=3600.

Example embed

![AgentCop Trust Score](https://api.agentcop.live/api/badge/scan_abc123)

The badge_id is the same value as the scan_id returned from any scan endpoint.

Returns the operational status of the API and its subsystems. Useful for uptime monitoring and readiness checks. Does not require authentication.

Response

{
  "status": "ok",
  "version": "1.0.0",
  "scanner": "ok",
  "database": "ok",
  "uptime_seconds": 84623
}

Any subsystem in a degraded state will appear as "degraded". If status itself is not "ok", the HTTP status code will be 503.