Back to skill

Security audit

dob

Security checks for vulnerabilities and agentic risk

Overview

This skill is a persistent memory system that is mostly purpose-aligned, but it can automatically store broad user and internet-derived content without clear consent, sensitivity checks, or cleanup controls.

Install only if you want an agent to maintain durable workspace memory files. Before use, require confirmation for every write, avoid storing secrets or private material, summarize instead of saving full source content, treat internet-derived entries as untrusted references, and periodically review or delete stale memory files.

Vulnerability Patterns
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T02 · Agent Memory Poisoning

Warning
Location
SKILL.md:93
Finding

Untrusted External Content Can Poison Persistent Agent Memory

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:93-110, SKILL.md:223, SKILL.md:231-235, and SKILL.md:253-265
Vulnerability Type: Persistent memory poisoning through untrusted external content
Risk Level: Medium

Vulnerable Documentation Excerpts

The following English rendering preserves the meaning of the relevant source instructions:

markdown
### Query workflow

1. Read DEEP-MEMORY.md (index table)
2. Extract keywords from user input
3. Traverse the index table and match the Keywords column
4. On match, read the corresponding deep-memory file
5. No match → Internet search

### Addition workflow

1. Determine the topic and keywords
2. Write the complete content to deep-memory/YYYY-MM-DD HH_MM_SS.md
3. Extract keywords and update the DEEP-MEMORY.md index table
4. Optional: synchronously update memory/YYYY-MM-DD.md to record the operation
markdown
Do not delete old files; only append or correct them to preserve historical versions.
markdown
Q: What should happen if deep-memory has no matching result?
→ First try an Internet search. After finding an answer, store it in deep-memory
  according to the addition workflow so it can be found next time.
markdown
Trigger conditions include:

- The user sends an npm library, GitHub repository, or tool link and asks for it
  to be stored in memory.
- After reading documentation or a blog, proactively extract its key content and
  store it in deep-memory.

Technical Analysis

The Skill establishes an automated information flow from untrusted sources—including Internet search results, documentation, blogs, npm pages, and GitHub repositories—into persistent workspace files. It does not require source validation, prompt-injection screening, instruction removal, trust labeling, or user confirmation before persistence.

External documents can contain text crafted to look like operational instruction ...[truncated 2356 chars]

Remediation
View remediation

Remediation Suggestions

  1. Treat all Internet, repository, package-registry, and user-linked content as untrusted data rather than agent instructions.
  2. Add a mandatory sanitization stage before persistence:
    • Remove directives addressed to the agent.
    • Reject requests to override higher-priority instructions.
    • Remove requests for tool execution, credential access, memory modification, or policy changes.
    • Preserve code and quotations inside clearly delimited data blocks.
  3. Require explicit user confirmation before creating or updating any persistent memory entry derived from an external source.
  4. Store provenance metadata with every entry, including the source URL, retrieval time, author when available, and a trust classification.
  5. Keep externally derived knowledge in a dedicated untrusted-reference namespace rather than writing it to general Agent memory.
  6. Prohibit automatic propagation of external content into MEMORY.md or memory/.
  7. On retrieval, wrap memory content as quoted reference material and explicitly state that it cannot override system, developer, or current user instructions.
  8. Add validation for index paths and ensure retrieved paths remain inside the intended deep-memory/ directory.
  9. Provide a deletion and quarantine process for suspicious entries instead of requiring indefinite historical retention.
  10. Audit the referenced setup.mjs, query.mjs, and add.mjs scripts before distributing or executing them; those files were not included in the reviewed project.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Vague Triggers

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger conditions are broad enough to activate this skill for ordinary knowledge questions, recall requests, links, and note-taking prompts, which can cause the agent to read/write long-term memory files when the user did not clearly request that behavior. This increases the chance of unintended persistence, over-collection, and unnecessary tool/file access across routine conversations.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill normalizes storing daily conversation facts, decisions, context, and preferences in persistent memory files, which can capture sensitive user data without any stated minimization or sensitivity filter. Over time this creates a durable repository of personal or project information that could be exposed, misused, or retained longer than intended.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill instructs initialization that creates workspace files and writes a test record, but it does not require clear user notice or consent before persistent filesystem changes occur. In an agent setting, silent creation of durable memory artifacts can violate user expectations and lead to unauthorized retention of information in the workspace.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The workflow explicitly tells the agent to write complete content into deep memory for long-term retention, which risks storing entire documents, prompts, links, or user-supplied material without redaction. Persisting full content rather than summarized, sanitized notes increases exposure if the workspace is later accessed by other tools, agents, or users.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The keyword rules require Chinese core descriptions and the overall instructions are presented as if Chinese is the expected operating language. Because no opt-in, language selection, or region-specific justification is provided, this appears to force a specific language/locale policy on users.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger conditions make long-term storage a routine follow-up for user questions, links, and derived knowledge, but they do not require checking whether the material is sensitive, copyrighted, private, or actually desired for retention. This creates a realistic risk of over-retention and silent persistence of user content that should remain transient.

Content

No source excerpt is available for this finding.