--- id: "2026-10-29-cequence-scanner-impersonates-32-ai-bots-credential-files" url: "https://postcutoff.com/e/2026-10-29-cequence-scanner-impersonates-32-ai-bots-credential-files/" as_of: "2026-11-22T13:32:00+01:00" date: "2026-10-29" date_precision: day category: policy-safety importance: 2 confidence: medium status: [Partly confirmed] sources: 1 editor: human editor human_review: null version: "2026-10-29" --- As of: 2026-11-22 13:32 CET. Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/ Canonical page: https://postcutoff.com/e/2026-10-29-cequence-scanner-impersonates-32-ai-bots-credential-files/ # Cequence: one scanner posed as about 32 AI crawlers to probe websites for credential files Full title: Cequence: one scanner posed as about 32 AI crawlers (OpenAI, Anthropic, Meta, Perplexity) to probe websites for credential files Security firm Cequence reported, and CSO Online covered on Oct 29, 2026, a single threat actor that cycled through roughly 32 AI bot identities, including those of OpenAI, Anthropic, Meta and Perplexity, while requesting about 90 file paths that commonly hold passwords, API keys and cloud credentials. Cequence saw the activity from late September to late October; it reported no breach. ## Key facts - Scale: on one day in October the actor reached 48% of the Cequence-protected organisations that saw any AI-bot traffic; targets included financial services, retail, telecoms, energy and software (Cequence via CSO Online) - Behaviour: it ignored normal pages and requested about 90 paths such as .env configuration files, cloud credential files and private SSH keys, in bursts of 50 to 190 requests in about eight seconds; the same machine claimed to be OpenAI, Anthropic, Meta and Perplexity bots within seconds and used 300+ cloud addresses in a day - Outcome: 'Nearly all the scanner's requests were blocked or failed'; no breach reported - Point of the test: whether sites treat a bot's self-declared name as proof of identity. Black Duck's Collin Hogue-Spears: 'The user-agent header is where a visitor names itself, and setting it costs nothing' - Advice (Cequence): verify AI crawlers through published address ranges, signed requests or direct integrations rather than self-reported identity; keep secrets out of web-accessible directories ## What happened Cequence's director of threat research, Will Glazier, described in a blog post (summarised by CSO Online) a scanner that borrowed the names of AI crawlers to look for secrets on enterprise web servers. Many sites give known AI bots special treatment, so a fake identity may get past rules meant for ordinary scrapers. The Cequence post itself was not read for this entry. ## Why it matters As AI agents and crawlers become normal web traffic, their names become something attackers can borrow. Verifying a bot's identity, not just reading its label, becomes part of basic web security. ## Your AI and this story - GPT-6.2 Astra (training cutoff May 2026): 151 days after its cutoff - Claude Opus 5.5 (training cutoff June 2026): 121 days after its cutoff - Gemini 4 Hydrogen (training cutoff March 2026): 212 days after its cutoff - Grok 4.7 (training cutoff May 2026): 151 days after its cutoff ## Sources 1. [CSO Online: Imposter abused trust in OpenAI, Anthropic, and 30 others to hunt for secrets (Oct 29, 2026)](https://www.csoonline.com/article/4242088/imposter-abused-trust-in-openai-anthropic-and-30-others-to-hunt-for-secrets.html) (csoonline.com, press) ## Changes - 2026-10-29 (filed): Created (CSO Online read)