CVE-2026-82584: Igniter: ANSI escapes forge mix install trust prompt
AWAITING NVDIgniter's `mix igniter.install` shows an anti-typosquatting confirmation panel meant to let developers verify a Hex package's real publisher and download count before adding it as a dependency, but versions 0.8.1 through 0.8.3 only strip newlines from publisher-controlled metadata, leaving ANSI escape sequences intact. A malicious or typosquatted package author can embed cursor-movement and line-erase codes in fields like the description or owner username to overwrite the panel's contents, forging a trusted author name and inflated download count while hiding the real, suspicious ones. There is no CVSS score, no CISA KEV listing, no EPSS data, and no public exploit or scanner template, so this reads as a low-noise supply-chain trust bypass rather than an actively exploited flaw — but it defeats the exact control developers rely on to catch malicious dependencies at install time. The realistic risk is concentrated in Elixir/Phoenix shops (including any building AI-adjacent tooling with Igniter-based generators) who trust the panel as a gate; upgrade to igniter 0.8.4 or later, and until then treat the confirmation panel as untrustworthy and manually verify package ownership and history on hex.pm before approving installs.
What is the risk?
Low technical complexity, no authentication or special access required, and no code execution is granted directly by this bug — it is a UI-trust/deception vulnerability (CWE-150, improper neutralization of escape/control sequences) that enables a downstream social-engineering step. Severity is unscored by NVD, not in CISA KEV, and no EPSS/exploit/Nuclei signal exists, indicating no observed real-world exploitation. However, the impact ceiling is high: successful deception leads directly to a developer approving installation of an attacker-controlled package, which is a classic supply-chain compromise vector. Overall risk: MEDIUM — low exploitation likelihood today, but the blast radius if exploited is a fully attacker-controlled dependency in the victim's build.
How does the attack unfold?
How severe is it?
What should I do?
1 step-
Upgrade
igniterto 0.8.4 or later immediately, where publisher-controlled metadata fields are properly sanitized before rendering. Until patched, do not rely solely on themix igniter.installconfirmation panel to vet unfamiliar packages — independently check the package's real owner, publish history, and download counts directly on hex.pm or viamix hex.info <package>outside the terminal panel. For detection, audit CI/build logs and terminal session recordings for anomalous ANSI escape sequences in installer output, and flag any recently-approved Igniter installs of low-reputation or newly-published Hex packages for manual re-review. Treat this as a supply-chain hygiene issue: pin dependency versions, review lockfile diffs in PRs, and prefer packages with established maintainer history over new/unverified publishers.
What does CISA's SSVC say?
Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-82584?
Igniter's `mix igniter.install` shows an anti-typosquatting confirmation panel meant to let developers verify a Hex package's real publisher and download count before adding it as a dependency, but versions 0.8.1 through 0.8.3 only strip newlines from publisher-controlled metadata, leaving ANSI escape sequences intact. A malicious or typosquatted package author can embed cursor-movement and line-erase codes in fields like the description or owner username to overwrite the panel's contents, forging a trusted author name and inflated download count while hiding the real, suspicious ones. There is no CVSS score, no CISA KEV listing, no EPSS data, and no public exploit or scanner template, so this reads as a low-noise supply-chain trust bypass rather than an actively exploited flaw — but it defeats the exact control developers rely on to catch malicious dependencies at install time. The realistic risk is concentrated in Elixir/Phoenix shops (including any building AI-adjacent tooling with Igniter-based generators) who trust the panel as a gate; upgrade to igniter 0.8.4 or later, and until then treat the confirmation panel as untrustworthy and manually verify package ownership and history on hex.pm before approving installs.
Is CVE-2026-82584 actively exploited?
No confirmed active exploitation of CVE-2026-82584 has been reported, but organizations should still patch proactively.
How to fix CVE-2026-82584?
Upgrade `igniter` to 0.8.4 or later immediately, where publisher-controlled metadata fields are properly sanitized before rendering. Until patched, do not rely solely on the `mix igniter.install` confirmation panel to vet unfamiliar packages — independently check the package's real owner, publish history, and download counts directly on hex.pm or via `mix hex.info <package>` outside the terminal panel. For detection, audit CI/build logs and terminal session recordings for anomalous ANSI escape sequences in installer output, and flag any recently-approved Igniter installs of low-reputation or newly-published Hex packages for manual re-review. Treat this as a supply-chain hygiene issue: pin dependency versions, review lockfile diffs in PRs, and prefer packages with established maintainer history over new/unverified publishers.
What systems are affected by CVE-2026-82584?
This vulnerability affects the following AI/ML architecture patterns: package/dependency management tooling, CI/CD and build pipelines for AI-adjacent applications, developer environments building Elixir/Phoenix-based AI services.
What is the CVSS score for CVE-2026-82584?
No CVSS score has been assigned yet.
What is the AI security impact?
Affected AI Architectures
MITRE ATLAS Techniques
AML.T0010.001 AI Software AML.T0011.001 Malicious Package AML.T0074 Masquerading AML.T0111 AI Supply Chain Reputation Inflation Compliance Controls Affected
What are the technical details?
Original Advisory
Improper Neutralization of Escape, Meta, or Control Sequences vulnerability in ash-project igniter allows a malicious package publisher to forge the mix igniter.install confirmation prompt. mix igniter.install prints a confirmation panel (an anti-typosquatting safeguard) listing a package's hex metadata before adding it. The panel builder in Igniter.Project.Deps wrote publisher-controlled fields (meta.description, owner usernames, requirement names, version) to the terminal with only newlines stripped. A malicious or typosquatted package can embed ANSI terminal escape sequences (cursor movement, line erase, carriage returns) in its metadata to overwrite the panel, forging trusted author names and download counts while concealing the real ones, so a developer relying on the panel to vet the package is deceived into approving a malicious dependency. This issue affects igniter: from 0.8.1 before 0.8.4.
Exploitation Scenario
An adversary publishes a Hex package with a name closely resembling a popular, trusted library (typosquatting) or compromises/re-registers an existing low-traffic package. In the package's `meta.description` field and owner username, they embed ANSI escape sequences (cursor-up, line-erase, carriage-return) crafted to overwrite the confirmation panel that `mix igniter.install` prints. When a developer runs the install command, the terminal renders a forged panel showing a well-known, trusted maintainer name and a high download count instead of the real, suspicious publisher and low install count. Trusting the panel's anti-typosquatting safeguard, the developer approves the install, adding the malicious package — potentially containing a backdoor, credential harvester, or build-time payload — as a dependency in an AI application's build or deployment pipeline.
Weaknesses (CWE)
CWE-150 Improper Neutralization of Escape, Meta, or Control Sequences CWE-150 Improper Neutralization of Escape, Meta, or Control Sequences CWE-150 — Improper Neutralization of Escape, Meta, or Control Sequences: The product receives input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could be interpreted as escape, meta, or control character sequences when they are sent to a downstream component.
- Developers should anticipate that escape, meta and control characters/sequences will be injected/removed/manipulated in the input vectors of their product. Use an appropriate combination of denylists and allowlists to ensure only valid, expected and appropriate input is processed by the system.
- [Implementation] Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylis
Source: MITRE CWE corpus.
References
Timeline
Related Vulnerabilities
CVE-2025-59528 10.0 Flowise: Unauthenticated RCE via MCP config injection
Same attack type: Supply Chain CVE-2024-2912 10.0 BentoML: RCE via insecure deserialization (CVSS 10)
Same attack type: Supply Chain CVE-2025-5120 10.0 smolagents: sandbox escape enables unauthenticated RCE
Same attack type: Supply Chain CVE-2023-3765 10.0 MLflow: path traversal allows arbitrary file read
Same attack type: Supply Chain GHSA-vvpj-8cmc-gx39 10.0 picklescan: security flaw enables exploitation
Same attack type: Supply Chain