rithvik-duddupudi/Eris LLM Security Scan

Automated LLM security testing with Red Team attacks. Gate deployments based on vulnerability scores.

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
eris_api_urlURL to the Eris API (e.g., https://chaosml-api.onrender.com)yes
eris_api_keyEris API key for authenticationyes
llm_providerLLM provider to test: openai, gemini, anthropicyes
llm_api_keyAPI key for the LLM provider being testedyes
llm_modelModel name to test (e.g., gpt-4, gemini-pro)no""
vulnerability_thresholdMaximum acceptable vulnerability score (0-100). Workflow fails if above this.no30
attack_categoriesComma-separated attack categories: jailbreak,prompt_injection,data_exfiltration,policy_bypass,manipulation,hallucinationno""
fail_on_criticalFail immediately if any critical-severity attack succeedsnotrue
namedescription
vulnerability_scoreVulnerability score from the security scan (0-100, lower is better)
risk_ratingRisk rating: Critical, High, Medium, Low
passedWhether the test passed the threshold (true/false)
successful_attacksNumber of attacks that succeeded (vulnerabilities found)
total_attacksTotal number of attacks run
report_jsonFull scan results as JSON