AgentProv: Auditing Agentic LLM API Providers via Tool-use Policy Probes

arXiv 2609.00052e65248abee9a128b447232ccddbf03ff9d46f72fba830143cc54e9300bf071fc
AI-governanceAutoGenCrewAIEEGLLM-securityLangGraphMCPagent-payment-protocolsagentic-ai-securityattack-forensicsauthorizationbearer-credentialsconfused-deputyformal-verificationindustrial-cybersecuritymodel-provenancemulti-agent-delegationprivacyprompt-injectionroboticsside-channel-analysissupply-chain-provenancetoken-replayx402zero-knowledge-proofs

Paper metadata

arXiv ID
2609.00052
Version
Not specified by this published record
Category
Computer Science — Cryptography and Security (cs.CR)

The PDF link points to arxiv.org. Baitaphish does not expose a private stored PDF.

Evidence and limitations

Source ID
arxiv_cs_cr
Record identifier
e65248abee9a128b447232ccddbf03ff9d46f72fba830143cc54e9300bf071fc
Enrichment time
2026-09-02T07:23:27Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.