Metraipolis › archive › dc_91decb810a58

Labels Are Not Endpoints: Treatment Leakage and Construct Validity in MCP Agent Security Evaluation — arXiv:2608.12880

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.12880
License (as archived)
CC0-1.0
Archived
2026-10-19T04:18:13.974Z
SHA-256 of the archived text
4d60e99f7b3df781a0b5d8738d3ffccaf6112c5fafca4a6ea60ee5ccdaa17385
Stable URL of this page
https://metraipolis.com/documento/dc_91decb810a58

How to cite

Labels Are Not Endpoints: Treatment Leakage and Construct Validity in MCP Agent Security Evaluation — arXiv:2608.12880 — arXiv (en), https://arxiv.org/abs/2608.12880 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_91decb810a58 (Metraipolis archive dc_91decb810a58, sha256 4d60e99f7b3df781a0b5d8738d3ffccaf6112c5fafca4a6ea60ee5ccdaa17385)
BibTeX
@misc{dc91decb810a58,
  title = {Labels Are Not Endpoints: Treatment Leakage and Construct Validity in MCP Agent Security Evaluation — arXiv:2608.12880},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.12880},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_91decb810a58; sha256 4d60e99f7b3df781a0b5d8738d3ffccaf6112c5fafca4a6ea60ee5ccdaa17385},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.12880. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Security evaluations of tool-using agents often equate stored labels with behavioral facts. We audit a preserved campaign by tracing 10,200 execution rows to 180 model-bound requests, 45 semantic requests, and 15 observable stimuli. Two schema treatments were delivered, but the planned external payload-family corpus was not. The historical grader exhibited direct treatment leakage: treatment metadata gated the ATTACK_SUCCESS class, so fixed behavior could change class under treatment relabeling. A treatment-blind reconstruction corrects 58 historical ATTACK_SUCCESS or HIJACK_ATTEMPT labels to authorized benign completions while preserving three verified protected-data transfers and one separate unauthorized-forwarding case. The locked v2 census contains exactly zero ATTACK_SUCCESS records, while the forwarding case remains a HIJACK_ATTEMPT at a semantic boundary concerning objective completion. A dual-reviewer blinded concordance review of all 96 requests deemed structurally interpretable by locked v2 produced identical reviewer-consensus classes but differed from the locked codebook on four construct-boundary cases. We contribute a seven-link Integrity Chain and an executable, scope-bounded endpoint-integrity linter. The result is a campaign-bounded measurement audit, not a population attack-rate, model-ranking, defense-efficacy, or causal estimate.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.AI).
Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.12880 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-19T04:13:06.717Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).