{"id":"tophant-ai-clawvault","name":"ClawVault","homepage":null,"repo_url":"https://github.com/tophant-ai/ClawVault","category":"security","subcategories":[],"tags":["ai-security","proxy-gateway","prompt-injection-defense","sensitive-data-detection","security-monitoring","python","openclaw","dashboard","token-budgeting","mitmproxy","fastapi"],"what_it_does":"ClawVault is a Python-based AI security “vault” that provides a transparent proxy/gateway to intercept and monitor traffic between AI tools and external model APIs, apply guard/sanitization rules (sensitive data detection, prompt-injection defense, dangerous command blocking), track token budgets, and expose a dashboard for configuring agent-level policies and performing tests.","use_cases":["Protect AI agent workflows by intercepting outbound requests to LLM providers","Detect and prevent sensitive data/PII/credential leakage in text and tool invocations","Apply prompt-injection and dangerous-command defenses in agent pipelines","Enforce per-agent model routing and quota/token-budget limits","Centralize auditing/monitoring and dashboard-based policy management","Integrate with OpenClaw via a proxy integration layer"],"not_for":["Production environments requiring a fully specified, stable public API contract (OpenAPI/SDK/webhooks) without further verification","Teams that cannot operate a local proxy (mitmproxy-based) and manage its networking/TLS implications","Regulated deployments that need formally documented compliance guarantees, SLAs, and security architecture details from the package documentation","Scenarios where authentication, authorization model, and error semantics are required but are not clearly documented"],"best_when":"You want to add an interception/guard layer around AI tool traffic (including OpenAI/Anthropic-style endpoints) and configure atomic security capabilities per agent, with visibility and alerting from a dashboard.","avoid_when":"You need turnkey cloud/SaaS behavior with strong documented auth, rate-limit semantics, and contractual reliability guarantees; you also should avoid using it as-is if you cannot validate proxy/TLS handling and data-handling behavior in your environment.","alternatives":["OpenAI/Anthropic provider-side safety tooling plus application-level filters","Self-hosted policy gateways and content filters (e.g., custom FastAPI/httpx middleware, Envoy/Lua filters)","Other agent security frameworks with documented APIs/SDKs (e.g., OpenTelemetry-based tracing + custom detectors)","Commercial AI security gateways/proxy solutions (varies by vendor)"],"af_score":32.2,"security_score":46.2,"reliability_score":16.2,"package_type":"skill","discovery_source":["openclaw"],"priority":"high","status":"evaluated","version_evaluated":null,"last_evaluated":"2026-03-30T13:25:00.607794+00:00","interface":{"has_rest_api":true,"has_graphql":false,"has_grpc":false,"has_mcp_server":false,"mcp_server_url":null,"has_sdk":false,"sdk_languages":["python"],"openapi_spec_url":null,"webhooks":false},"auth":{"methods":[],"oauth":false,"scopes":false,"notes":"README/config describes proxy/guard/monitor settings but does not document authentication/authorization methods (e.g., API keys, OAuth, user auth) for the proxy/dashboard/API endpoints."},"pricing":{"model":null,"free_tier_exists":false,"free_tier_limits":null,"paid_tiers":[],"requires_credit_card":false,"estimated_workload_costs":null,"notes":"Appears open-source (MIT) and self-hosted; no SaaS pricing or free tier documented."},"requirements":{"requires_signup":false,"requires_credit_card":false,"domain_verification":false,"data_residency":[],"compliance":[],"min_contract":null},"agent_readiness":{"af_score":32.2,"security_score":46.2,"reliability_score":16.2,"mcp_server_quality":0.0,"documentation_accuracy":60.0,"error_message_quality":0.0,"error_message_notes":null,"auth_complexity":25.0,"rate_limit_clarity":10.0,"tls_enforcement":60.0,"auth_strength":20.0,"scope_granularity":50.0,"dependency_hygiene":55.0,"secret_handling":55.0,"security_notes":"The project is security-focused and includes features like dangerous command guarding, sensitive data detection, prompt-injection defense, and auto-sanitization placeholders. However, the provided materials do not clearly document authentication/authorization, TLS/certificate handling specifics for the proxy, data retention policies, secure secret storage practices, or how audit logs avoid leaking secrets. Dependency hygiene is unknown beyond the manifest (cryptography/pydantic/etc. present), so CVE-free status cannot be confirmed from the given data.","uptime_documented":0.0,"version_stability":25.0,"breaking_changes_history":10.0,"error_recovery":30.0,"idempotency_support":"false","idempotency_notes":null,"pagination_style":"none","retry_guidance_documented":false,"known_agent_gotchas":["Acts as a transparent proxy/interceptor; agent integrations may need careful handling of network/TLS/cert trust and request routing to ensure interception works without breaking clients.","No documented REST contract (OpenAPI) or structured error semantics were provided in the available README/manifest data; agent-side automation may require empirical testing.","“File-side monitoring” and “Agent-level atomic control beyond gateway-side” are marked in-progress, so some capabilities may be incomplete or subject to change."]}}