Back to skill

Security audit

OMEM memory

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed self-hosted memory skill, with setup risks users should manage around the external server package, persistent memory, and endpoint security.

Install only if you want persistent agent memory on an OMEM server you control. Prefer a local loopback server or HTTPS for any remote server, pin and review the omem-infrastructure package before installation, and do not store credentials or sensitive one-off data in memory.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Findings (2)

T08 · Insecure Dependencies

Warning
Location
skill.md:35
Finding

Unpinned Third-Party Package Installation

Content
View full analysis
Remediation
View remediation
" ``` 2. Distribute a requirements or lock file containing cryptographic hashes and install with hash verification: ```bash python -m pip install --require-hashes -r requirements.txt ``` 3. Pin and audit all transitive dependencies rather than controlling only the top-level package version. 4. State the expected official package index and upstream source repository so users can identify dependency-confusion or package-substitution attempts. 5. Use an isolated virtual environment and avoid installing or running the server with administrative or root privileges. 6. Establish a documented update process in which new versions and their dependency changes are reviewed before the pinned version is advanced. 7. Update both `skill.md` and `README.md` so that neither document continues to recommend an unpinned installation. ]]>

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/omem.py:28
Finding

Bearer Credentials and Memory Data Can Be Sent Over Plaintext HTTP

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (3)

Tainted flow: 'r' from os.environ.get (line 39, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Confidence
90% confidence
Finding

Credentials or environment variables flow to a network sink. This is a high-confidence indicator of credential exfiltration.

Content

Scanner excerpt · scripts/omem.py (reported line 43)May include surrounding context.

python
"Content-Type": "application/json",
        "Authorization": f"Bearer {KEY}"})
    try:
        with urllib.request.urlopen(r, timeout=30) as resp:
            return json.load(resp)
    except urllib.error.HTTPError as e:
        if tolerate:

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · skill.md (reported line 72)May include surrounding context.

text

The state is one of BELIEVED_TRUE, BELIEVED_FALSE, CONTRADICTED, UNKNOWN.
Treat CONTRADICTED as "do not act on this without asking the user": the
record holds conflicting information from different sources, and OMEM
deliberately refuses to pick a winner for you.