PYSEC-2026-3666
Vulnerability from pysec - Published: 2026-08-19 11:56 - Updated: 2026-08-19 12:16Summary
CVE-2026-32608 ("Command Injection via Process Names in Action Command Templates") was fixed (commit 5680a5d) by adding _sanitize_mustache_dict, which replaces the shell operators &&, |, >>, > with spaces in the values rendered into action command templates.
The sanitizer only processes top-level string values (if isinstance(v, str)). Attacker-controlled nested values — most notably a process's cmdline, which Glances exposes as a list and which is fully attacker-controlled via argv — are passed through unsanitized. Because the Mustache renderer (chevron) does not HTML-escape the pipe character |, a | embedded in such a nested value survives into the rendered command and is then interpreted by secure_popen (which still interprets &&/|/> by default, allow_operators=True), re-introducing the exact command injection the CVE was meant to close.
Details
The fix (glances/actions.py):
_SHELL_OPERATORS = ('&&', '|', '>>', '>') # line 25
def _sanitize_mustache_dict(mustache_dict): # line 28
...
for k, v in mustache_dict.items():
if isinstance(v, str): # line 40 <-- ONLY top-level strings
for op in _SHELL_OPERATORS:
v = v.replace(op, ' ')
safe[k] = v
else:
safe[k] = v # nested list/dict passed VERBATIM
return safe
Render + sink (glances/actions.py:104-111):
safe_dict = _sanitize_mustache_dict(mustache_dict)
cmd_full = chevron.render(cmd, safe_dict) # chevron does NOT escape '|'
...
ret = secure_popen(cmd_full) # secure_popen(cmd, allow_operators=True)
secure_popen (glances/secure.py:17, default allow_operators=True) splits the command by &&, then __secure_popen interprets | (pipe to a new process) and > (write output to a file). A surviving | therefore launches an attacker-named second process.
The attacker-controlled nested value — cmdline. The action mustache_dict is the per-item plugin stat (glances/plugins/plugin/model.py:931 mustache_dict = item, then :943 self.actions.run(..., mustache_dict=mustache_dict)). For the processlist plugin, each item contains cmdline, a list of the process arguments, set by the attacker simply by launching a process with chosen argv. The sanitizer's isinstance(v, str) test skips the list, so its elements reach chevron.render unmodified.
Why the operator survives render. chevron/Mustache HTML-escapes & < > " ' for {{var}} (so > and && are neutralized) but does not escape |. A pipe in the (unsanitized) nested value therefore reaches secure_popen intact and is interpreted.
Parent-fix attribution (verified against the real diff of commit 5680a5d / CVE-2026-32608): that fix added exactly _SHELL_OPERATORS, _sanitize_mustache_dict, and the _sanitize_mustache_dict(mustache_dict) call — and the sanitizer's docstring explicitly claims to neutralize "user-controllable data (process names, container names, mount points, etc.)". It does so only for top-level strings; the list/dict case (else: safe[k] = v) was left unsanitized. This is therefore a genuine incomplete-fix gap, not a re-report of the patched (top-level string) vector.
Proof of Concept
Lab-only, harmless (touches a marker file; non-destructive). Runs the real glances chain (_sanitize_mustache_dict → chevron.render → secure_popen) — see poc/glances_nested_mustache_poc.py.
Attacker process argv (the only attacker input): cmdline = ['x', '|touch /tmp/glances_poc_marker', '#'].
Admin action template (renders the offending process's cmdline): echo ALERT {{#cmdline}}{{.}} {{/cmdline}}.
Observed (confirmed on develop HEAD 92156d0/4.5.6 and verified code-identical on v4.5.5):
cmdline after sanitizer : ['x', '|touch /tmp/glances_poc_marker', '#'] <- pipe survives
cmd_full -> secure_popen : 'echo ALERT x |touch /tmp/glances_poc_marker # '
[VULNERABLE] marker created -> /tmp/glances_poc_marker (command injection executed)
Replacing touch /tmp/... with any command yields arbitrary execution in the Glances process context.
Preconditions (stated honestly)
- A configured alert action whose command template renders a nested stat field (e.g. the process
cmdlinevia a{{#cmdline}}…{{/cmdline}}section). Templates that render only flat string fields ({{name}},{{value}},{{username}},{{mnt_point}}) are not affected — those values are sanitized. - Glances running with privilege to enumerate the attacker's process (typically root in server/agent monitoring deployments) → privilege boundary crossed (
S:C).
Impact
A local unprivileged user gains OS command execution in the Glances security context (commonly root) — the same impact and threat model as the parent CVE-2026-32608, re-enabled for any action template that renders a nested stat field. The injection is reliable once the (admin-set) template references such a field.
Suggested fix
- Sanitize recursively — apply the operator stripping to strings inside lists and dicts, not only top-level
strvalues. - And/or build the templated action as an argument list and run it via
secure_popen(..., allow_operators=False)/shell=Falsewithout operator interpretation. - And/or also strip the pipe
|(and treat all_SHELL_OPERATORS) on every rendered string regardless of nesting; do not rely on Mustache HTML-escaping (it does not escape|).
Credit
Ta Duc Thien
| Name | purl | glances | pkg:pypi/glances |
|---|
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "glances",
"purl": "pkg:pypi/glances"
},
"ranges": [
{
"events": [
{
"introduced": "4.5.2"
},
{
"fixed": "4.5.6"
}
],
"type": "ECOSYSTEM"
}
],
"versions": [
"4.5.2",
"4.5.3",
"4.5.4",
"4.5.5"
]
}
],
"aliases": [
"CVE-2026-62982",
"GHSA-73wf-9vmv-5pv9"
],
"details": "## Summary\nCVE-2026-32608 (\"Command Injection via Process Names in Action Command Templates\") was fixed (commit `5680a5d`) by adding `_sanitize_mustache_dict`, which replaces the shell operators `\u0026\u0026`, `|`, `\u003e\u003e`, `\u003e` with spaces in the values rendered into action command templates.\n\nThe sanitizer only processes **top-level string** values (`if isinstance(v, str)`). Attacker-controlled **nested** values \u2014 most notably a process\u0027s **`cmdline`, which Glances exposes as a `list`** and which is fully attacker-controlled via argv \u2014 are passed through **unsanitized**. Because the Mustache renderer (`chevron`) does **not** HTML-escape the pipe character `|`, a `|` embedded in such a nested value survives into the rendered command and is then interpreted by `secure_popen` (which still interprets `\u0026\u0026`/`|`/`\u003e` by default, `allow_operators=True`), re-introducing the exact command injection the CVE was meant to close.\n\n## Details\nThe fix (`glances/actions.py`):\n```python\n_SHELL_OPERATORS = (\u0027\u0026\u0026\u0027, \u0027|\u0027, \u0027\u003e\u003e\u0027, \u0027\u003e\u0027) # line 25\n\ndef _sanitize_mustache_dict(mustache_dict): # line 28\n ...\n for k, v in mustache_dict.items():\n if isinstance(v, str): # line 40 \u003c-- ONLY top-level strings\n for op in _SHELL_OPERATORS:\n v = v.replace(op, \u0027 \u0027)\n safe[k] = v\n else:\n safe[k] = v # nested list/dict passed VERBATIM\n return safe\n```\nRender + sink (`glances/actions.py:104-111`):\n```python\nsafe_dict = _sanitize_mustache_dict(mustache_dict)\ncmd_full = chevron.render(cmd, safe_dict) # chevron does NOT escape \u0027|\u0027\n...\nret = secure_popen(cmd_full) # secure_popen(cmd, allow_operators=True)\n```\n`secure_popen` (`glances/secure.py:17`, default `allow_operators=True`) splits the command by `\u0026\u0026`, then `__secure_popen` interprets `|` (pipe to a new process) and `\u003e` (write output to a file). A surviving `|` therefore launches an attacker-named second process.\n\n**The attacker-controlled nested value \u2014 `cmdline`.** The action `mustache_dict` is the per-item plugin stat (`glances/plugins/plugin/model.py:931` `mustache_dict = item`, then `:943` `self.actions.run(..., mustache_dict=mustache_dict)`). For the processlist plugin, each `item` contains `cmdline`, a **list** of the process arguments, set by the attacker simply by launching a process with chosen argv. The sanitizer\u0027s `isinstance(v, str)` test skips the list, so its elements reach `chevron.render` unmodified.\n\n**Why the operator survives render.** `chevron`/Mustache HTML-escapes `\u0026 \u003c \u003e \" \u0027` for `{{var}}` (so `\u003e` and `\u0026\u0026` are neutralized) but **does not escape `|`**. A pipe in the (unsanitized) nested value therefore reaches `secure_popen` intact and is interpreted.\n\n**Parent-fix attribution (verified against the real diff of commit `5680a5d` / CVE-2026-32608):** that fix added exactly `_SHELL_OPERATORS`, `_sanitize_mustache_dict`, and the `_sanitize_mustache_dict(mustache_dict)` call \u2014 and the sanitizer\u0027s docstring explicitly claims to neutralize \"user-controllable data (process names, container names, mount points, etc.)\". It does so only for top-level strings; the list/dict case (`else: safe[k] = v`) was left unsanitized. This is therefore a genuine incomplete-fix gap, not a re-report of the patched (top-level string) vector.\n\n## Proof of Concept\nLab-only, harmless (touches a marker file; non-destructive). Runs the real `glances` chain (`_sanitize_mustache_dict` \u2192 `chevron.render` \u2192 `secure_popen`) \u2014 see `poc/glances_nested_mustache_poc.py`.\n\nAttacker process argv (the only attacker input): `cmdline = [\u0027x\u0027, \u0027|touch /tmp/glances_poc_marker\u0027, \u0027#\u0027]`.\nAdmin action template (renders the offending process\u0027s cmdline): `echo ALERT {{#cmdline}}{{.}} {{/cmdline}}`.\n\nObserved (confirmed on develop HEAD `92156d0`/4.5.6 and verified code-identical on v4.5.5):\n```\ncmdline after sanitizer : [\u0027x\u0027, \u0027|touch /tmp/glances_poc_marker\u0027, \u0027#\u0027] \u003c- pipe survives\ncmd_full -\u003e secure_popen : \u0027echo ALERT x |touch /tmp/glances_poc_marker # \u0027\n[VULNERABLE] marker created -\u003e /tmp/glances_poc_marker (command injection executed)\n```\nReplacing `touch /tmp/...` with any command yields arbitrary execution in the Glances process context.\n\n## Preconditions (stated honestly)\n- A configured alert **action** whose command template **renders a nested stat field** (e.g. the process `cmdline` via a `{{#cmdline}}\u2026{{/cmdline}}` section). Templates that render only **flat string** fields (`{{name}}`, `{{value}}`, `{{username}}`, `{{mnt_point}}`) are **not** affected \u2014 those values *are* sanitized.\n- Glances running with privilege to enumerate the attacker\u0027s process (typically **root** in server/agent monitoring deployments) \u2192 privilege boundary crossed (`S:C`).\n\n## Impact\nA local unprivileged user gains OS command execution in the Glances security context (commonly root) \u2014 the same impact and threat model as the parent CVE-2026-32608, re-enabled for any action template that renders a nested stat field. The injection is reliable once the (admin-set) template references such a field.\n\n## Suggested fix\n- Sanitize **recursively** \u2014 apply the operator stripping to strings inside lists and dicts, not only top-level `str` values.\n- And/or build the templated action as an argument list and run it via `secure_popen(..., allow_operators=False)` / `shell=False` without operator interpretation.\n- And/or also strip the pipe `|` (and treat all `_SHELL_OPERATORS`) on every rendered string regardless of nesting; do not rely on Mustache HTML-escaping (it does not escape `|`).\n\n## Credit\nTa Duc Thien",
"id": "PYSEC-2026-3666",
"modified": "2026-08-19T12:16:24.926132Z",
"published": "2026-08-19T11:56:26.862162Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/nicolargo/glances/security/advisories/GHSA-73wf-9vmv-5pv9"
},
{
"type": "WEB",
"url": "https://github.com/nicolargo/glances/commit/ea4cf2f54f0d961e24aa0b24fff9584bab39db93"
},
{
"type": "PACKAGE",
"url": "https://github.com/nicolargo/glances"
},
{
"type": "WEB",
"url": "https://github.com/nicolargo/glances/releases/tag/v4.5.6"
},
{
"type": "PACKAGE",
"url": "https://pypi.org/project/glances"
},
{
"type": "ADVISORY",
"url": "https://github.com/advisories/GHSA-73wf-9vmv-5pv9"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-62982"
}
],
"severity": [
{
"score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H",
"type": "CVSS_V3"
}
],
"summary": "Glances: Incomplete fix of CVE-2026-32608: action-template sanitizer is bypassed by nested stat values (process \u0027cmdline\u0027) \u2192 OS command injection"
}
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.
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.