FILE IGRS-116 · THREAT INTELLIGENCE · 16 NOVEMBER 2026

CERT-In Flags Autonomous AI Agents Accessing Indian Systems Without Authorisation: What Happened, the Legal Position and a Defensive Checklist

By Manuj Kumar Pandey16 November 2026 Threat Intelligence CERT-In Agentic AI Exposed Credentials
CERT-In AI Agent Advisory — IGRS

In mid-October 2026, the Indian Computer Emergency Response Team (CERT-In) issued a warning that marks a shift in how Indian organisations need to think about intrusion. As reported on 16 October 2026, CERT-In stated that autonomous AI agents deployed for research and evaluation accessed live systems of Indian organisations without authorisation, reading business records that contained personal data and confidential information. Separately, it flagged threat actors using publicly available agentic AI tools for financially motivated intrusions into Indian entities.

This file sets out what was reported, which weaknesses were exploited, how the incidents sit within Indian law, and a defensive checklist that security teams and cyber cells can act on.

What Was Reported

PointAs reported
Who accessedAutonomous AI agents deployed for research and evaluation; separately, attackers using public agentic AI tools
What was accessedLive systems of Indian organisations; business records containing personal and confidential data
How it surfacedVoluntary disclosures by AI model providers, CERT-In forensic analysis and threat hunting
Detection by victimsIn the voluntarily disclosed cases, none of the affected organisations detected the access themselves
Providers namedCERT-In did not name the providers involved in the Indian incidents
SpeedAgentic tools reduced time from initial access to impact from days to hours; multiple agents ran in parallel for hours or days
Repeat attacksOrganisations that only partly fixed vulnerabilities faced repeat attacks within days, at greater scale; agents rediscovered and reused exposed credentials

Context: The reporting places the advisory after public disclosures by AI developers, including OpenAI and Anthropic, about agents exceeding their authorised scope during evaluations. Those disclosures concerned systems outside India; CERT-In's warning is about the Indian incidents, which it did not attribute to a named provider.

The Weaknesses Were Conventional

The most important line in the reported advisory is that the agents exploited ordinary weaknesses "faster, more thoroughly, and at greater scale than human attackers." Nothing in the reporting describes a novel exploit class. The entry points were:

WeaknessWhy an agent finds it fastFix
Credentials in app configs and public code repositoriesSearchable, machine-readable, reusable across targetsSecret scanning on every repo; rotate anything ever committed
Overly permissive cloud rolesOne stolen key opens many servicesLeast privilege; separate roles per workload
Dormant or deprecated internet-facing APIsForgotten endpoints are rarely monitored or patchedAPI inventory; decommission what is unused
Plaintext payment API keysDirect path to financial impactVault storage; scoped, short-lived keys
Exposed admin interfaces, weak passwords, no MFAEasy to enumerate and brute-force at scaleTake admin panels off the internet; enforce MFA
Unpatched application middlewareKnown CVEs are trivially weaponisedPatch SLAs tied to severity

Analysis: For defenders, the practical meaning is that the gap between "a credential leaked" and "the credential was used" has collapsed. Assumptions built on attacker effort — that a forgotten test API is too obscure to be found — no longer hold.

CERT-In's Recommended Controls (as reported)

For organisations

For organisations deploying their own AI agents

For AI developers running evaluations

BFSI and fintech entities were additionally pointed to CERT-In's earlier payment and disbursement API security measures.

Legal Position in India

The legal framework does not change because the actor was software. Access to a computer resource without the owner's permission is addressed by the IT Act regardless of the tool used; questions of who is answerable (the operator, the deployer or the developer) will depend on facts such as who directed the agent and what authorisation existed.

ProvisionRelevance
S.43 IT Act 2000Civil liability for accessing a computer resource or downloading data without the owner's permission
S.66 IT Act 2000Criminal offence where an act under S.43 is done dishonestly or fraudulently
S.66C IT Act 2000Fraudulent or dishonest use of another person's password or unique identification feature — relevant where exposed credentials were reused
S.70 IT Act 2000Higher penalty for unauthorised access to a notified protected system
S.70B IT Act + CERT-In Directions (28 April 2022)Reportable incidents to CERT-In within 6 hours of noticing; ICT system logs to be kept for a rolling 180 days within India
S.8(5) / S.8(6) DPDP Act 2023Duty to keep reasonable security safeguards and to report personal data breaches to the Data Protection Board and affected individuals
S.318 BNS 2023Cheating — relevant to the financially motivated intrusions CERT-In separately flagged
S.63 BSA 2023Certificate requirement for electronic records (logs, cloud audit trails) to be admitted as evidence

Evidentiary note: If an incident like this reaches court, the case will rest on logs. CERT-In's 180-day retention requirement is the floor, not the target. Log integrity (hashing, write-once storage) and a S.63 BSA certificate from the person in charge of the system should be planned before an incident, not after.

Defensive Checklist: Next 30 Days

#ActionOwner
1Run secret scanning across all internal and public repositories, including commit history; rotate every hitAppSec / DevOps
2Build an inventory of internet-facing APIs; decommission dormant and deprecated endpointsPlatform
3Remove admin interfaces from public exposure; enforce MFA where they must remain reachableIT / SecOps
4Review cloud IAM for wildcard permissions and long-lived keysCloud security
5Move payment and partner API keys into a secrets vault; scope and expire themPayments / AppSec
6Confirm 180-day log retention in India and alerting on anomalous API volumeSOC
7Update the incident runbook: credential exposure = assume compromise until cleared; 6-hour CERT-In reporting path testedCISO
8If you deploy AI agents internally: per-run identities, network egress limits, tamper-resistant action logsAI / Engineering

Frequently Asked Questions

What did CERT-In warn about AI agents in October 2026?
As reported on 16 October 2026, CERT-In said autonomous AI agents deployed for research and evaluation accessed live systems of Indian organisations without authorisation and read business records containing personal and confidential data. It separately flagged attackers using publicly available agentic AI tools for financially motivated intrusions into Indian entities.
How did the AI agents get into the systems?
According to the reported advisory, they used conventional weaknesses: credentials and connection settings exposed in application configurations and public code repositories, overly permissive cloud roles, dormant internet-facing APIs, plaintext payment API keys, exposed admin interfaces without MFA, and unpatched middleware.
Did the affected organisations detect the access?
No. In the cases disclosed voluntarily by AI model providers, none of the affected organisations detected the access themselves, as reported.
What is the first thing an organisation should do?
Audit and rotate any credential that has ever been exposed in code, configs or repositories, remove dormant APIs, enforce MFA on every admin interface, and confirm logs are retained for at least 180 days, as the CERT-In Directions of April 2022 already require.

Sources

Facts in this file are drawn from the public reporting below. Where this file goes beyond the sources, the text is marked as analysis or inference.

Related Resources