IETF individual draft for __usage_ai-training robots.txt signal
Draft not yet an RFC; creates a usage-purpose User-Agent token within existing robots.txt syntax, independent of ordinary crawling directives.
AI security research / sources
Primary reports, advisories, papers and original records behind the Claims. Each record supports specific evidence rather than a whole Case by default.
1209 Evidence · Page 5 of 41
30 matching records on this page
Draft not yet an RFC; creates a usage-purpose User-Agent token within existing robots.txt syntax, independent of ordinary crawling directives.
First-party paper on instrumented AI-serving and MCP honeypots and attempted credential extraction using encoded requests.
IBM Granite model post-training experiments distinguishing instruction-bearing and lower-trust data tokens.
Advisory on optional Slack signing secret verification and the resulting message/file retrieval boundary.
AgentDyn controlled tests of resource-scoping against indirect injection, with explicit legitimate-task regression.
Original study of role laundering from quoted third-party dialogue into an agent’s apparent user-preference memory.
Original study auditing public skills and examining security effects of composing otherwise ordinary-looking skills.
Experimental isolation of lower-trust content and prompt-injection resistance on AgentDojo tests.
Independent comparison of Open VSX append-only registry event history, npm metadata and Cursor marketplace incident confirmation.
Observed invoice-fraud email with hidden Unicode Tag characters intended to redirect an AI email assistant.
Research describing tool-output text that impersonates user or conversation roles and its measured effect on agent behavior.
Controlled comparisons of identical prompt payloads placed in repository instructions, source code, or tool data.
JFrog paper proposing signed process attestations that link agent sessions, model/tool/harness identity, approvals, and created artifacts.
Agno issue documenting model-selected pandas function dispatch and the read_pickle deserialization risk.
Report on internal token-use incentives and management reminder that AI usage does not define engineering output.
On-chain reconstruction with a separate detection vendor and a commercial-product disclosure; counts and categorization distinguished from Binance reporting.
Independent replay of sixty-four Qwen continuations testing whether a blocked outbound call had a stable attacker destination.
Talos report on AI API access in command-and-control and separate coercive modules in the same malware tooling.
Paper testing conditional agent misbehavior after small pretraining poisoning and subsequent clean supervised or preference training.
CrowdStrike telemetry of an operator-controlled ARTEX campaign, with distinct target versus verified post-exploitation populations.
Forever Security reports local-file exposure, privileged browser bridge message spoofing and an origin display spoof across several desktop assistants; conditions differ by product.
Epoch study of GLM-5.3 assisted chain construction from a pending Firefox renderer patch and a separately discovered file-protection bypass, in controlled versions.
GitHub issue #30014 was opened October 29, 2026. The reporter traces function calls emitted from a failed streaming turn, their persistence through a Retry event, duplicate execution, and orphaned tool responses. The issue is not proof of an independently confirmed production-loss rate or a fixed version.
Open research artifact indexing hidden injection messages in GitHub content and noting false positives and attack-attempt denominators.
Independent Kubeflow Pipelines authorization or artifact handling disclosure.
Independent Kubeflow Pipelines authorization or artifact handling disclosure.
Microsoft analysis of distinct North Korean activity clusters, tool adoption, counterfeit recruiting, and limitations of observed AI involvement.
NVIDIA first-party bulletin covering separate missing-authentication and SQL-injection vulnerabilities in toolkit versions.
OpenAI October 29 first-party description of L0-L3 agent monitoring and the distinctions between detection, investigation, paging and interruption.
Original analysis of the Load/Mine campaign, executable model configuration, mining infrastructure and privileged GPUStack variant.
No matches on this page.