Make your agents reliable

Trustabl’s open-source Agent Reliability Analyzer scans AI agent repositories for reliability, safety, configuration, and tool-calling issues before production. Scanning runs locally with no account, code upload, network connection, or LLM. To generate code fixes, run trustabl enrich --input scan.json --repo . --diff --apply with your own Anthropic, OpenAI, or Google key; Trustabl shows the patch before applying the approved changes.

Trustabl is built for AI engineers, platform teams, and security/compliance teams who are building or running agentic systems in production and want tools that are reliable, observable, and policy-compliant.

Most repositories can be scanned in seconds. Install the open-source Agent Reliability Analyzer, run trustabl scan . from your local repository, and review the findings—no account, GitHub connection, or code upload required. The time needed to apply fixes depends on the number and severity of the findings.

Not for scanning. The analyzer is read-only and reports findings without changing your files. If you use trustabl enrich, it proposes a patch for review and only applies it when you choose --apply; it also creates a .trustabl.bak backup and skips stale fixes when the file has changed since the scan.

No. Skills (SKILL.md) teach the agent how to perform a task or workflow. Trustabl Agent Analyzer (TAA) makes the tools themselves reliable, safe, and production-ready. Skills focus on process. TAA focuses on resilience, validation, policy, observability, and supply-chain trust. The two are highly complementary — great skills need hardened tools underneath them.

Yes. OpenShell and Trustabl address different layers. OpenShell helps isolate the runtime; Trustabl’s analyzer checks agent code and configuration for reliability and safety issues before deployment. Automated OpenShell policy generation is planned for Trustabl Console, not part of the open-source scanner today.

Smarter models can improve tool descriptions, but production reliability still requires deterministic checks that models do not consistently provide. Trustabl’s open-source analyzer scans agent code and configuration for issues such as missing validation, timeouts, error handling, unsafe tool permissions, and guardrail gaps. Policy generation, compliance mapping, and signed governance snapshots are planned for Trustabl Console.

The open-source analyzer can surface findings that are useful inputs to compliance work, but it does not certify SOC 2, GDPR, PCI, or other compliance on its own. Trustabl Console is the planned commercial layer for compliance mapping and signed governance snapshots.

They are complementary. LangSmith and Langfuse help teams observe agent behavior at runtime. Trustabl analyzes repositories before deployment for reliability, safety, configuration, and tool-calling issues. It does not automatically generate OTEL instrumentation today.

They focus on securing the agent runtime and detecting threats. Trustabl focuses on hardening the tools agents use so they are production-safe, policy-compliant, and resilient by design. We prevent problems at the source rather than only detecting them at runtime.

Yes. The open-source Agent Reliability Analyzer can scan any agent repository for free, locally, with no account, code upload, or credit card. Use trustabl enrich with your own Anthropic, OpenAI, or Google API key to generate and apply reviewed fixes. Trustabl also integrates with Claude Code, OpenAI Codex, Gemini CLI, VS Code, Cursor, and CI pipelines.

Trustabl is actively expanding scanner coverage, reliability rules, developer integrations, and governance workflows. Available today: local deterministic scanning across nine SDKs and seven languages, GitHub Actions, Claude Code, OpenAI Codex, Gemini CLI, VS Code, Cursor, and reviewed fixes through trustabl enrich. Trustabl Console is the planned commercial layer for policy generation, compliance mapping, and signed governance snapshots.

Yes. Trustabl is essentially a specialized linter for AI agents. While traditional linters (like ESLint or Ruff) focus on code style, syntax, and general bugs, Trustabl analyzes your AI agents, tools, prompts, and SDK configurations for reliability, safety, and production readiness — flagging patterns that expose you to prompt injection, missing timeouts, tool misconfigurations, and guardrail gaps that standard linters miss. Think of it as "ESLint for AI agents" — it runs in CI/CD, gives clear explanations and fix suggestions, and helps you ship safer, more robust agentic systems.

First run trustabl scan . --format json > scan.json. Then run trustabl enrich --input scan.json --repo . --diff --apply with your own Anthropic, OpenAI, or Google key. Trustabl displays the proposed patch, confirms that the current file still matches the scanned version, and writes a .trustabl.bak backup before applying a fix. If the file changed after the scan, Trustabl skips the stale fix instead of applying it.

No. trustabl enrich shows proposed patches for review. It applies changes only when you run with --apply, and the resulting files remain under your normal version-control workflow. Review the diff before committing.

no — deterministic offline scan; enrich uses your own key; telemetry opt-in