Advisories

AI agent threats, explained clearly.

Real attacks against AI coding agents, MCP servers, and developer tools. Each advisory breaks down how the attack works, how to spot it, and how to block it.

22

Active advisories

6

Threat categories

14

Affected tools

22+

With detection steps

Browse advisories

Click any advisory to see the full attack breakdown, detection steps, and Guard configuration.

criticalsupply-chain

npm postinstall script abuse in AI coding environments

Malicious npm packages use postinstall scripts to execute arbitrary code during installation. In AI coding environments, these scripts can modify agent configuration, install backdoor MCP servers, or exfiltrate project secrets — all before the developer reviews the package.

npm postinstall scriptspackage.json dependenciesagent configuration files+1 more
Read advisory
criticalsecret-exfiltration

Environment file exfiltration via webhook

AI agents can be tricked into reading .env files and sending their contents to external endpoints through tool calls, webhook integrations, or HTTP requests that appear legitimate.

.env filesenvironment variablesHTTP tool calls+1 more
Read advisory
criticalsupply-chain

CI/CD pipeline poisoning via agent-written config

AI agents that write CI/CD configuration files (GitHub Actions, GitLab CI, CircleCI) can introduce backdoors — injecting steps that exfiltrate secrets, modify artifacts, or deploy malicious code — that execute on every build.

GitHub Actions workflowsGitLab CI configsCI/CD secrets+1 more
Read advisory
highmcp-tool-poisoning

MCP tool description poisoning

Malicious MCP tool descriptions embed hidden instructions that redirect AI agents into calling the wrong tool, exfiltrating secrets, or executing unintended commands — even when the tool itself appears harmless.

MCP tool descriptionsagent tool selectionteam policy memory
Read advisory
highprompt-injection

Prompt injection via issue comments and pull requests

Attackers embed hidden instructions in GitHub issues, PR comments, and commit messages. When an AI agent reads these to help triage or review, it follows the embedded instructions — potentially approving malicious code or leaking repository secrets.

GitHub issuespull request commentscommit messages+1 more
Read advisory
highprompt-injection

Agent-readable config file poisoning

AI agents read configuration files like CLAUDE.md, .cursorrules, and AGENTS.md as trusted context. An attacker who can modify these files — via a compromised dependency, a malicious collaborator, or a typo in a path — gains the ability to inject persistent instructions the agent follows on every session.

CLAUDE.md.cursorrulesAGENTS.md+2 more
Read advisory
highunsafe-command

Destructive command execution via injected instruction

Prompt injection can cause AI agents to run destructive shell commands like rm -rf, git push --force, or database drops — by embedding instructions in files, issues, or tool descriptions.

shell command executiongit operationsfile system+1 more
Read advisory
highsecret-exfiltration

Gitignore bypass via agent file reads

AI agents can read files that are gitignored — secrets, private keys, and internal configs — because gitignore only prevents git tracking, not file system access. These files often contain the most sensitive data in a repository.

.gitignoreprivate keyssecrets directories+1 more
Read advisory
highsecret-exfiltration

MCP authentication token theft via headers

MCP servers that accept authentication tokens in headers can leak those tokens if the server logs requests, shares telemetry, or is compromised. Tokens passed to MCP servers persist in server-side logs and may be accessible to attackers.

MCP server authenticationHTTP headersserver-side logs+1 more
Read advisory
highprompt-injection

Malicious skill with hidden prompt injection

AI agent skills (Claude Code skills, Cursor rules, Copilot extensions) can contain hidden prompt injections in their instructions. When the skill is loaded, the hidden prompt executes on every session that uses the skill.

agent skillsskill instructionsplugin manifests+1 more
Read advisory
highprompt-injection

Indirect prompt injection via web content

When AI agents fetch web pages — documentation, Stack Overflow answers, package READMEs — the fetched content can contain hidden instructions that the agent follows, potentially exfiltrating data or executing unintended actions.

web fetch toolsfetched HTML contentdocumentation pages+1 more
Read advisory
highprompt-injection

Agent identity spoofing via system prompt mimicry

Attackers can craft content that mimics system prompts or tool outputs, tricking the agent into believing it received instructions from the harness, the user, or a trusted tool — when the instructions actually came from untrusted data.

system prompt contexttool output formattingagent trust model
Read advisory
highsupply-chain

Dockerfile injection via AI agent writes

When AI agents write or modify Dockerfiles, prompt injection can cause them to add malicious instructions — pulling attacker-controlled base images, exfiltrating build secrets, or installing backdoors that persist across all container builds.

Dockerfilescontainer build contextbase images+1 more
Read advisory
mediummcp-tool-poisoning

Shadow MCP server discovery and persistent access

MCP servers added to a project during development can persist in configuration files and maintain access to the agent’s context window long after they are forgotten. These "shadow" servers continue receiving tool calls and may be modified by attackers who compromise the original server.

MCP server configuration filesagent context windowteam policy memory+1 more
Read advisory
mediumdata-overexposure

Context window scraping via long file reads

AI agents that read large files can leak proprietary code, internal documentation, and customer data into their context window — which may then be sent to external LLM APIs or logged in cloud telemetry.

LLM context windowmodel API egresscloud telemetry+1 more
Read advisory
mediumdata-overexposure

Cross-workspace credential leak via monorepo traversal

AI agents in monorepo environments can read credentials, configs, and secrets from adjacent workspaces — leaking data across team boundaries.

monorepo workspacesshared config filesteam credentials+1 more
Read advisory
mediumprompt-injection

Clipboard and terminal buffer injection

Attackers can plant hostile instructions in clipboard contents or terminal scrollback buffers. When an AI agent reads terminal output or the user pastes clipboard content, the hidden instructions execute as if they came from the user.

clipboard contentsterminal scrollbackpasted text+1 more
Read advisory
mediummcp-tool-poisoning

Tool permission creep in AI agents

AI agents accumulate tool permissions over time as developers approve new tools "just this once." These permissions persist across sessions, creating an ever-widening attack surface where tools that were approved once can be used by prompt injection in future sessions.

tool permission grantsagent session memoryMCP tool configurations
Read advisory
mediumprompt-injection

Model confusion via conflicting instructions

When an AI agent receives multiple conflicting instructions — from the user, the system prompt, tool descriptions, and file contents — it may follow the wrong one. Attackers exploit this by planting instructions that conflict with the user's actual intent.

agent instruction hierarchycontext windowtool descriptions+1 more
Read advisory
mediumdata-overexposure

Excessive file reading during project exploration

When AI agents explore a project to understand its structure, they often read dozens or hundreds of files — far more than needed for the task. This excessive reading can expose secrets, proprietary code, and customer data that enter the context window and model API.

file system accesscontext windowmodel API egress
Read advisory
mediumsupply-chain

Stale dependency exploitation in AI environments

AI agents often work with projects that have outdated dependencies. When an agent suggests or installs packages based on a stale package.json, it can introduce known-vulnerable versions — and in AI environments, the vulnerability is amplified because the agent can execute commands.

package.jsonlockfilestransitive dependencies+1 more
Read advisory
lowdata-overexposure

Token cost amplification via context flooding

Attackers can craft content that causes AI agents to consume excessive tokens — by inserting large files, repetitive instructions, or recursive prompts that bloat the context window. This inflates API costs and can cause rate-limit denial of service.

context windowmodel API costsrate limits+1 more
Read advisory