Reworkd — agentic threat model
Reworkd presents a high agentic risk profile due to its capability to dynamically generate and execute code for web scraping and self-healing. The primary threat vector is indirect prompt injection from untrusted web pages, which could exploit the code generation and execution pipeline.
OWASP AIVSS score rationale
| Autonomy of Action | 0.80 | |
| Goal-Driven Planning | 0.80 | |
| Self-Modification | 0.70 | |
| Dynamic Tool Use | 0.80 | |
| Persistent Memory | 0.40 | |
| Contextual Awareness | 0.70 | |
| Dynamic Identity | 0.30 | |
| Multi-Agent Interactions | 0.60 | |
| Non-Determinism | 0.70 | |
| Opacity & Reflexivity | 0.60 |
Scored with the canonical OWASP AIVSS formula (AIVSS calculator reference); agentic risk factors estimated from the agent’s described capabilities.
MAESTRO 7-layer threat model
Per-layer threats for this agent. Layers tagged “not certain from listing” are general, caveated commentary where the public description didn’t pin that layer.
Not certain from the listing — The specific foundation models used are not disclosed. However, the model is highly vulnerable to indirect prompt injection and adversarial examples embedded in target websites, which could manipulate the code generation process.
The agent ingests untrusted external web data dynamically. This introduces severe risks of data poisoning, indirect prompt injection via scraped content, and potential data exfiltration if the agent is tricked into sending scraped data to unauthorized endpoints.
The framework orchestrates scanning, code generation, execution, and self-healing. The primary risk is insecure tool integration and tool misuse, specifically the execution of dynamically generated scraping code which could contain malicious payloads.
Not certain from the listing — The execution environment for the generated scraping code is not described. If the environment lacks strict sandboxing, container isolation, or egress filtering, it is highly vulnerable to remote code execution (RCE) and lateral network movement.
The agent features automated validation of results and self-healing scrapers. However, there is a risk of validation bypass or evaluation gaming where malicious web structures trick the validation agent into accepting poisoned or incomplete data.
Not certain from the listing — No specific compliance certifications (e.g., SOC2, ISO 27001), access control policies, or audit logging mechanisms are detailed in the public directory listing.
The platform utilizes multiple specialized agents (scanning, generation, execution, validation). This creates a risk of cascading failures or trust abuse, where a compromised scanning or generation agent successfully deceives the execution or validation agents.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology. Are you the vendor? Factual corrections are free.