GHSA-9P44-J4G5-CFX5
Vulnerability from github – Published: 2026-02-18 15:24 – Updated: 2026-02-19 21:56Command Injection in aquasecurity/trivy-action via Unsanitized Environment Variable Export
A command injection vulnerability exists in aquasecurity/trivy-action due to improper handling of action inputs when exporting environment variables. The action writes export VAR=<input> lines to trivy_envs.txt based on user-supplied inputs and subsequently sources this file in entrypoint.sh.
Because input values are written without appropriate shell escaping, attacker-controlled input containing shell metacharacters (e.g., $(...), backticks, or other command substitution syntax) may be evaluated during the sourcing process. This can result in arbitrary command execution within the GitHub Actions runner context.
Severity:
Moderate
CVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:U/C:H/I:H/A:N
CWE-78: Improper Neutralization of Special Elements used in an OS Command (‘OS Command Injection’)
Impact:
Successful exploitation may lead to arbitrary command execution in the CI runner environment.
Affected Versions:
- Versions >= 0.31.0 and <= 0.33.1
- Introduced in commit
7aca5ac
Affected Conditions:
The vulnerability is exploitable when a consuming workflow passes attacker-controlled data into any action input that is written to trivy_envs.txt. Access to user input is required by the malicious actor.
A representative exploitation pattern involves incorporating untrusted pull request metadata into an action parameter. For example:
- uses: aquasecurity/trivy-action@0.33.1
with:
output: "trivy-${{ github.event.pull_request.title }}.sarif"
If the pull request title contains shell syntax, it may be executed when the generated environment file is sourced.
Not Affected:
- Workflows that do not pass attacker-controlled data into
trivy-actioninputs - Workflows that upgrade to a patched version that properly escapes shell values or eliminates the
source ./trivy_envs.txtpattern - Workflows where user input is not accessible.
Call Sites:
action.yaml:188—set_env_var_if_providedwrites unescapedexportlinesentrypoint.sh:9— sources./trivy_envs.txt
{
"affected": [
{
"package": {
"ecosystem": "GitHub Actions",
"name": "aquasecurity/trivy-action"
},
"ranges": [
{
"events": [
{
"introduced": "0.31.0"
},
{
"fixed": "0.34.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-26189"
],
"database_specific": {
"cwe_ids": [
"CWE-78"
],
"github_reviewed": true,
"github_reviewed_at": "2026-02-18T15:24:43Z",
"nvd_published_at": "2026-02-19T20:25:42Z",
"severity": "MODERATE"
},
"details": "Command Injection in aquasecurity/trivy-action via Unsanitized Environment Variable Export\n\n\nA command injection vulnerability exists in `aquasecurity/trivy-action` due to improper handling of action inputs when exporting environment variables. The action writes `export VAR=\u003cinput\u003e` lines to `trivy_envs.txt` based on user-supplied inputs and subsequently sources this file in `entrypoint.sh`.\n\nBecause input values are written without appropriate shell escaping, attacker-controlled input containing shell metacharacters (e.g., `$(...)`, backticks, or other command substitution syntax) may be evaluated during the sourcing process. This can result in arbitrary command execution within the GitHub Actions runner context.\n\n**Severity:**\n\nModerate\n\nCVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:U/C:H/I:H/A:N\n\nCWE-78: Improper Neutralization of Special Elements used in an OS Command (\u2018OS Command Injection\u2019)\n\n**Impact:**\n\nSuccessful exploitation may lead to arbitrary command execution in the CI runner environment.\n\n\n**Affected Versions:**\n\n* Versions \u003e= 0.31.0 and \u003c= 0.33.1\n* Introduced in commit `7aca5ac`\n\n**Affected Conditions:**\n\nThe vulnerability is exploitable when a consuming workflow passes attacker-controlled data into any action input that is written to `trivy_envs.txt`. Access to user input is required by the malicious actor.\n\nA representative exploitation pattern involves incorporating untrusted pull request metadata into an action parameter. For example:\n\n```yaml\n- uses: aquasecurity/trivy-action@0.33.1\n with:\n output: \"trivy-${{ github.event.pull_request.title }}.sarif\"\n```\n\nIf the pull request title contains shell syntax, it may be executed when the generated environment file is sourced.\n\n**Not Affected:**\n\n* Workflows that do not pass attacker-controlled data into `trivy-action` inputs\n* Workflows that upgrade to a patched version that properly escapes shell values or eliminates the `source ./trivy_envs.txt` pattern\n* Workflows where user input is not accessible.\n\n**Call Sites:**\n\n* `action.yaml:188` \u2014 `set_env_var_if_provided` writes unescaped `export` lines\n* `entrypoint.sh:9` \u2014 sources `./trivy_envs.txt`",
"id": "GHSA-9p44-j4g5-cfx5",
"modified": "2026-02-19T21:56:21Z",
"published": "2026-02-18T15:24:43Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/aquasecurity/trivy-action/security/advisories/GHSA-9p44-j4g5-cfx5"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-26189"
},
{
"type": "WEB",
"url": "https://github.com/aquasecurity/trivy-action/commit/7aca5acc9500b463826cc47a47a65ad7d404b045"
},
{
"type": "WEB",
"url": "https://github.com/aquasecurity/trivy-action/commit/bc61dc55704e2d5704760f3cdab0d09acf16e4ca"
},
{
"type": "PACKAGE",
"url": "https://github.com/aquasecurity/trivy-action"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:U/C:H/I:H/A:N",
"type": "CVSS_V3"
}
],
"summary": "Trivy Action has a script injection via sourced env file in composite action"
}
Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
Vulnerability timeline
Sightings and recorded dates, grouped by day · UTC
How durations are calculated
Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.
Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.
Loading timeline…
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.