protologue

Excessive Agency

Excessive agency is the vulnerability of giving an LLM-based system more functionality, permissions, or autonomy than its task requires, so that a manipulated or mistaken model can take damaging actions.

Description

It is listed as LLM06 in the OWASP Top 10 for LLM Applications (2025), which names three root causes, excessive functionality, excessive permissions, and excessive autonomy. Mitigations are least-privilege tools and credentials, narrowly scoped functions instead of open-ended ones, and human approval for high-impact actions.

Sources

  1. OWASP Gen AI Security Project (2025). LLM06:2025 Excessive Agency.

Cite this entry

Protologue. (2026). Excessive Agency. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0, PTL-0166). https://protologue.com/t/excessive-agency/

BibTeX
@misc{protologue_excessive_agency,
  title = {Excessive Agency},
  author = {{Protologue}},
  year = {2026},
  howpublished = {Protologue: A Taxonomy of Prompting and LLM Techniques, v1.4.0},
  note = {Entry PTL-0166},
  url = {https://protologue.com/t/excessive-agency/}
}

Markdown JSON