The authors built a multi-step autoformalization pipeline that translates a procedural manual into Linear Temporal Logic, a formal language for expressing requirements about ordering and future actions.
Runtime shields guide LLM agents through rule-bound workflows more reliably
MENTOR converts natural-language manuals into context-aware constraints and guidance, improving compliance across eight workflow domains.
Research Lab
Ashwani Anand · Ivi Chatzi · Ritam Raha · Anne-Kathrin Schmuck
Max Planck Institute for Software Systems · TNO-ESI
Research Digest··3 min read
Anand and colleagues present MENTOR, a runtime shield that checks an agent’s proposed tool calls while also recommending compliant ways to proceed.
Why this paper
From Max Planck Institute for Software Systems and TNO-ESI
In one line
MENTOR auto-formalizes procedural manuals to block non-compliant tool calls and guide LLM agents toward compliant actions.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§