Home/ Incident database/litellm-pypi-supply-chain-2026-03
Data breach Credential compromise verified

Compromised LiteLLM PyPI packages harvest credentials via supply-chain attack

LiteLLM (BerriAI) · Mar 24, 2026 · Coding agent

What happened

Malicious versions of the litellm Python package (1.82.7 and 1.82.8) were published to PyPI using compromised maintainer credentials, part of the TeamPCP campaign. The payload harvested SSH keys, AWS/GCP/Azure credentials, Kubernetes tokens and database passwords, then exfiltrated them to an attacker-controlled domain. The packages were live for roughly 40 minutes before PyPI quarantined them.

Loss / impact
Mar 24
2026
Credential
Failure mode
Coding
Agent type

Causal vector

Supply-chain compromise of an LLM gateway package — malicious code executed on install and harvested developer/CI credentials

Source

Reported by LiteLLM (security update). Verified against the primary report.

Read the original report ↗

How a spend firewall would have helped

This is the supply-chain face of the same problem: agent tooling expands the credential blast radius. A spend firewall's value is the policy boundary — limiting what any single agent or tool can touch and spend, so a compromised dependency cannot silently monetize your infrastructure.

The six rule types that contain this class of failure

Per-transaction cap

Any single spend above your ceiling is BLOCKED before it moves.

Daily total

Cumulative spend across all agent calls, bounded per day.

Velocity limit

Stops runaway retry loops — the #1 cause of overnight losses.

Merchant allowlist

Only approved destinations can ever receive funds.

Category rules

Flag high-risk classes (crypto, infra, refunds) for review.

Approval threshold

Above a value, the action waits for a human.

Related incidents

Don't be the next entry

Every incident in this database is the result of trusting a prompt, a provider cap, or a human review cycle. sipi.bot replaces all three with one deterministic call. 67 documented failures, one control.