Action Firewall - Block Destructive AI Agent Actions

Gate agent-generated shell commands with deterministic allowlists and checks for injection, destructive operations, privilege escalation and download-to-shell patterns.

AI Agent Tool Firewall

zentrafoundry/zentra-agentops-tool-firewall

Allow, deny, redact, or require review for AI-agent tool calls.

Parse SQL and enforce a deterministic single-statement read-only policy before agent execution.

Action.com Search Scraper

powerai/action-com-search-scraper

Scrape products from Action.com by providing a search URL, with automatic pagination and comprehensive product information extraction.

Evaluate proposed MCP tool calls against deterministic allow, deny, domain and confirmation policies.

AI Agent Effect Firewall

tmas_from_iceland/ai-agent-effect-firewall

Execution firewall for AI agents. Blocks untrusted inputs from triggering real-world actions by enforcing decision-level allow / deny / approval gates. Infrastructure safety layer for tool-enabled agents

OFAC Recent Actions as they are posted: the action's own identifier, its date, its headline and its category (Sanctions List Updates, Regulations and Guidance, and so on). ACTION-LISTING SCOPE: this watches the published actions themselves.

Next Action Agent

westerly_palisade/next-action-by-dravara

Choose exactly one action from a supplied list. No 11-step plans. Read the default dataset. Set maxTotalChargeUsd above $0.03.

Action Scraper

appealing_jingle/action-scraper

Scrapes product data from Action.com including title, description, price, price description, category, URL, and image. Supports multiple languages, categories, sorting, pagination, and Apify proxy. Exports structured results to dataset (JSON, CSV).

Extract action items, owners, due hints and priorities from meeting transcripts or messy notes.