agent-ops-check
A free and open-source Node CLI static risk scanning tool for detecting operational risks in AI agents (data leakage, irreversible operations, hidden instructions and similar issues) during code review. It runs fully offline and supports a regex rule engine plus optional LLM review.
Timeline (last 90 days)
agent-ops-check is a free, MIT-licensed Node CLI tool that reads agent code, prompts, and configuration files and outputs a risk report, running entirely offline with no SaaS, no account, and no network requests.
agent-ops-check can be installed and used with the one-line command 'npx github:ultimatixarup/agent-ops-check ./my-agent', and its input supports folders, single files, .zip archives, or stdin.
The rules of agent-ops-check are written in YAML, the engine is based on regex matching, the report supports four formats—HTML, JSON, Markdown, and SARIF—and the local UI only listens on 127.0.0.1.
The LLM review in agent-ops-check is optional and uses the user's own API Key, while the static rules engine can complete the scan independently without connecting to any model.
agent-ops-check summarizes agent operational risks into five directions: Crossover (personal data), Actions (site operations), Governance/Labs (governance and overhead), ADSF (document-hidden instructions), and STE (unclear instruction phrasing).
agent-ops-check scores on a range from 0 to 100, deducting 10 points for high-risk findings, 4 points for medium-risk, and 1 point for low-risk, and the repository's own bad example scores 27.
agent-ops-check's checks are mapped to 112 items across 10 frameworks, including OWASP Top 10 for LLM Applications 2025, OWASP Agentic 2026, NIST AI RMF, ISO/IEC 42001, MITRE ATLAS, CSA AICM and the EU AI Act.
agent-ops-check clearly distinguishes 'maps to' from 'compliant'; the tool only does mapping, not compliance claims.
The author re-rated the 112 items based on the code's actual capabilities: 29 ✅, 64 🟡, 19 ⬜, versus the published status of 43/52/17, with 16 items rated by the author lower than the coverage status the tool publicly claims.
agent-ops-check's gate detection only recognizes a gate when it appears between 6 lines above and 12 lines below a given action, and guard logic outside that window is ignored, causing misjudgments.
4 more timeline events
All Facts (14)
agent-ops-check is a free, MIT-licensed Node CLI tool that reads agent code, prompts, and configuration files and outputs a risk report, running entirely offline with no SaaS, no account, and no network requests.
50%Unverifiedagent-ops-check can be installed and used with the one-line command 'npx github:ultimatixarup/agent-ops-check ./my-agent', and its input supports folders, single files, .zip archives, or stdin.
50%UnverifiedThe rules of agent-ops-check are written in YAML, the engine is based on regex matching, the report supports four formats—HTML, JSON, Markdown, and SARIF—and the local UI only listens on 127.0.0.1.
50%UnverifiedThe LLM review in agent-ops-check is optional and uses the user's own API Key, while the static rules engine can complete the scan independently without connecting to any model.
50%Unverifiedagent-ops-check summarizes agent operational risks into five directions: Crossover (personal data), Actions (site operations), Governance/Labs (governance and overhead), ADSF (document-hidden instructions), and STE (unclear instruction phrasing).
50%Unverifiedagent-ops-check scores on a range from 0 to 100, deducting 10 points for high-risk findings, 4 points for medium-risk, and 1 point for low-risk, and the repository's own bad example scores 27.
50%Unverifiedagent-ops-check's checks are mapped to 112 items across 10 frameworks, including OWASP Top 10 for LLM Applications 2025, OWASP Agentic 2026, NIST AI RMF, ISO/IEC 42001, MITRE ATLAS, CSA AICM and the EU AI Act.
50%Unverifiedagent-ops-check clearly distinguishes 'maps to' from 'compliant'; the tool only does mapping, not compliance claims.
50%UnverifiedThe author re-rated the 112 items based on the code's actual capabilities: 29 ✅, 64 🟡, 19 ⬜, versus the published status of 43/52/17, with 16 items rated by the author lower than the coverage status the tool publicly claims.
50%Unverifiedagent-ops-check's gate detection only recognizes a gate when it appears between 6 lines above and 12 lines below a given action, and guard logic outside that window is ignored, causing misjudgments.
50%Unverifiedagent-ops-check is a static checking tool that can only see what the code 'says' and cannot read actual runtime behavior, producing both false negatives and false positives.
50%Unverifiedagent-ops-check's static detection cannot cover all variants of prompt injection, so it still needs to be used in combination with runtime defenses.
50%UnverifiedThe author of the agent-ops-check project states that this is an independent project completed on personal time and unrelated to their employer.
50%UnverifiedAfter agent-ops-check outputs SARIF format, it can be seamlessly integrated into existing code review pipelines, and risk findings appear in the Pull Request diff view just like ESLint or Semgrep alerts.
50%