Severity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N
create_doc is a local/self-hosted MCP tool needing client access (AV:L, PR:L); traversal enables arbitrary file write (I:H) with only incidental read/overwrite (C:L), not the NVD's confidentiality-only model.
Primary rating from Vendor (mitre).
CVSS VectorVendor: mitre
Lifecycle Timeline
5DescriptionCVE.org
Directory traversal vulnerability in knowns-dev/knowns 0.11.4 via crafted folder name value to the create_doc tool.
AnalysisAI
Directory traversal in Knowns 0.11.4 (the 'knowns' npm CLI/MCP knowledge layer for AI-assisted development) lets an attacker supply a crafted folder name to the create_doc tool and cause documents to be created outside the intended workspace directory. Because create_doc is exposed to AI agents via MCP, a malicious or manipulated tool call can traverse the filesystem using sequences such as '../'. A researcher gist and an upstream fix commit are referenced, so publicly available exploit code exists; the flaw is not in CISA KEV and carries a low EPSS score (0.26%, 18th percentile), indicating no evidence of widespread active exploitation.
Technical ContextAI
Knowns is a CLI-first, self-hostable 'memory layer' for AI-native development that stores structured project context (specs, tasks, docs) and exposes it to AI assistants through the Model Context Protocol (MCP). One of those MCP tools, create_doc, accepts a folder name that is used to build the on-disk path for a new document. The root cause is CWE-22 (Improper Limitation of a Pathname to a Restricted Directory): the folder-name input is not canonicalized or constrained to the intended documents root, so path separators and parent-directory sequences ('../') in the value are honored, allowing writes outside the project store. Note that the CPE in the feed is a placeholder (cpe:2.3:a:n/a:n/a:*:*:*), so the affected component must be taken from the description (knowns-dev/knowns 0.11.4) rather than from structured CPE data.
RemediationAI
Upstream fix available (PR/commit); released patched version not independently confirmed - the referenced commit 62fb0daaa0b9e66fce1d42c02b5bb519367cf669 is cited as the fix, but the diff provided in the feed only shows README changes, so operators should track knowns-dev/knowns for the tagged npm release that supersedes 0.11.4 and upgrade once published (verify at https://www.npmjs.com/package/knowns). Until a confirmed patched version is installed, restrict who can reach create_doc: do not expose the Knowns MCP server or Web UI to untrusted networks, bind it to localhost, and place it behind authentication if remote access is needed (trade-off: breaks remote/team MCP usage). As a compensating control, run the Knowns process under a low-privilege account inside a container or sandbox whose only writable mount is the intended docs directory, so a traversed path cannot escape the workspace (trade-off: added deployment complexity). Treat create_doc arguments coming from AI agents as untrusted and review or constrain the folder-name value before execution.
LibreNMS before 24.10.0 allows a remote attacker to execute arbitrary code via OS command injection involving AboutContr
Reachable assertion in Open5GS up to version 2.7.6 allows remote unauthenticated attackers to crash the Diameter CER han
A Command Injection issue in the payload build page in BYOB (Build Your Own Botnet) 2.0 allows attackers to execute arbi
Denial of service in Open5GS 2.7.6 and earlier allows network-adjacent attackers with low-privilege Diameter Gx peer acc
Denial of service in Open5GS HSS (Home Subscriber Server) versions 2.7.0 through 2.7.6 is triggered remotely by manipula
Remote command injection in Control Web Panel allows unauthenticated attackers to execute arbitrary OS commands as root
LINQPad before 5.52.01 Pro edition is vulnerable to Unsafe Deserialization in LINQPad.AutoRefManager::PopulateFromCache(
Remote code execution in dash-uploader (Python package for Plotly Dash) versions 0.1.0 through 0.7.0a2 allows unauthenti
Command injection in LiteLLM 1.18.10 lets attackers who can configure MCP (Model Context Protocol) servers supply arbitr
SQL injection in AzuraCast's backup restore flow (versions up to and including 0.23.7) lets an attacker achieve privileg
Arbitrary file read in Yu AI Code Mother v4.3 lets anonymous, remote attackers retrieve files outside the intended previ
Remote code execution in APScheduler (all versions through 3.10.x and 4.0.0a5) is achievable when applications deseriali
Same weakness CWE-22 – Path Traversal
View allSame technique Path Traversal
View allShare
External POC / Exploit Code
Leaving vuln.today