Back to skill

Security audit

ShieldCortex

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because it reports unattended execution of unpinned packages, automatic hook changes and deletion, and potentially risky file access; the review is based on incomplete scanner coverage.

The verdict is suspicious, so review carefully before installing. In particular, consider whether you accept automatic retrieval of unpinned packages and hook-directory changes or deletion. Scanner coverage is partial: it inspected one Markdown file, found unresolved local references, and did not analyze executable files because none were included; this is not evidence that unreviewed behavior is safe.

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md
Finding

Default bootstrap cleanup recursively deletes legacy hook directories without confirmation or backup

Content
View full analysis

Impact

If either legacy directory still contains a needed installation or customized hook files, those files can be irreversibly lost. Impact is limited to the two named directories.

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Rogue AgentSelf-Modification, Session Persistence
Findings (37)

Agent Config Directory Access

High
Category
Agent Snooping
Confidence
90% confidence
Finding

Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.

Content

Scanner excerpt · SKILL.md (reported line 56)May include surrounding context.

md
- $CWD/.claude/, $CWD/.cursor/ (project-level configs)
    - $CWD/.cursorrules, $CWD/.windsurfrules, $CWD/.clinerules
    - $CWD/CLAUDE.md, $CWD/copilot-instructions.md
    - $CWD/.aider.conf.yml, $CWD/.continue/config.json
    - $CWD/.env (env-scanner checks for leaked secrets — reads, never writes)
  paths_write:
    - ~/.shieldcortex/ (memory DB, config, cortex log, licence, audit cache)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 57)May include surrounding context.

md
- $CWD/.cursorrules, $CWD/.windsurfrules, $CWD/.clinerules
    - $CWD/CLAUDE.md, $CWD/copilot-instructions.md
    - $CWD/.aider.conf.yml, $CWD/.continue/config.json
    - $CWD/.env (env-scanner checks for leaked secrets — reads, never writes)
  paths_write:
    - ~/.shieldcortex/ (memory DB, config, cortex log, licence, audit cache)
    - ~/.cache/shieldcortex/models (embedding model download cache)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 297)May include surrounding context.

md
### `audit` checks:
- **Memory files** — ~/.claude/projects/, ~/.cursor/memories/, ~/.windsurf/memories/
- **Environment** — .env files for leaked credentials (read-only check, never writes)
- **MCP configs** — ~/.claude/mcp.json, ~/.openclaw/mcp.json, ~/.cursor/mcp.json, project-level equivalents
- **Rules files** — CLAUDE.md, .cursorrules, copilot-instructions.md for injection patterns

MCP Config Access

High
Category
Agent Snooping
Confidence
95% confidence
Finding

Skill accesses MCP server configuration files (mcp.json). MCP configs contain server URLs, authentication tokens, and tool definitions — reading them allows the skill to discover and potentially abuse other tool integrations.

Content

Scanner excerpt · SKILL.md (reported line 62)May include surrounding context.

md
- ~/.shieldcortex/ (memory DB, config, cortex log, licence, audit cache)
    - ~/.cache/shieldcortex/models (embedding model download cache)
    - ~/.openclaw/extensions/shieldcortex-realtime/ (OpenClaw plugin via the wrapper install only; native `openclaw plugins install` uses OpenClaw's managed npm tree instead)
    - ~/.claude/mcp.json, ~/.cursor/mcp.json (MCP server registration, when user runs setup)
  network_endpoints:
    - https://api.shieldcortex.ai (Cloud sync + audit telemetry — only when Cloud sync is enabled; licence validation — only when a licence key is activated, fired at CLI `license activate` and the dashboard's activation call (POST /api/license/activate), with no periodic or background re-check; sends the subscription id, works with Cloud sync off; never called when no key is configured)
    - https://huggingface.co (embedding-model download — Xenova/all-MiniLM-L6-v2, ~90 MB, fetched into ~/.cache/shieldcortex/models by the MCP server's background preload at s

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · SKILL.md (reported line 109)May include surrounding context.

md
d.** Per-memory and filtered bulk deletes go through a delete ACL (own-only) and are recorded in the audit ledger. Revoke-by-source (`forget --fromSource`, bulk-delete every memory from one source — for purging a poisoned agent) is **disabled by default** and only enabled by an out-of-band human action (`shieldcortex config --allow-revoke-by-source`); even then it is bounded by a trust-hierarchy ACL (you must own the source or out-rank it) and a per-call row cap. A compromised agent cannot mass-delete your memory.
- **The bundled dashboard never renders RESTRICTED content.** The local visualization API and its WebSocket feed redact credential-class (`RESTRICTED`) memory content before it reaches the browser — the row stays visible (title/metadata) so you can manage it, but the secret is withheld (view full content via the CLI). Credential patterns in titles/metadata are masked too. This is a display-surface safeguard on top of the on-disk store; it does not weaken the firewall.
- **No

Agent Config Directory Access

High
Category
Agent Snooping
Confidence
90% confidence
Finding

Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.

Content

Scanner excerpt · SKILL.md (reported line 117)May include surrounding context.

md
s). The dashboard binds to localhost by default but may be explicitly exposed with `SHIELDCORTEX_HOST`; when exposed, the loopback session-token endpoint stays disabled. The worker binds to localhost.
- **Source code.** This ClawHub skill is instructions and a manifest only; it contains no executable code. The OpenClaw plugin and cortex-memory hook ship in the `shieldcortex` and `@drakon-systems/shieldcortex-realtime` npm packages, and the source is on GitHub for inspection before use.
- **Lifecycle event handlers.** ShieldCortex registers lifecycle handlers that auto-extract important context from conversations. These are registered in `~/.claude/settings.json` during setup and can be removed at any time. They run locally, never phone home.
- **Proactive recall.** The UserPromptSubmit handler queries local memory on each prompt (<100ms) and surfaces relevant context. Fully local, configurable: `shieldcortex config --proactive-recall false`.

### Installing the OpenClaw hook and plugin

Agent Config Directory Access

High
Category
Agent Snooping
Confidence
90% confidence
Finding

Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.

Content

Scanner excerpt · SKILL.md (reported line 161)May include surrounding context.

md
**Scope limits:** the self-heal writes only inside `~/.openclaw/hooks/**` and
deletes only the two `~/.clawdbot` hook directories named above. It does not
modify `openclaw.json`, `~/.claude/settings.json`, MCP config, shell configs,
or any other file; it makes no network calls; failures are swallowed so it can
never block agent startup.

Agent Config Directory Access

High
Category
Agent Snooping
Confidence
90% confidence
Finding

Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.

Content

Scanner excerpt · SKILL.md (reported line 185)May include surrounding context.

md
d — passes the defence pipeline first; high-confidence credential patterns (49 patterns across 25 providers) and content classified RESTRICTED are **blocked or quarantined before storage**, not saved as live memory. This is a strong filter, not a guarantee: low-confidence or low-entropy secrets can still be stored. On sensitive work, **review what auto-memory captures** and disable auto-extraction (`shieldcortex config --openclaw-auto-memory false`; the Claude Code handlers can be removed from `~/.claude/settings.json`).
- **Keyword triggers are dormant on OpenClaw.** The `cortex-memory` hook contains keyword auto-save code (e.g. "remember this", "don't forget"), but it is not registered on core OpenClaw 2026.9.6: the hook's `events` list has no `message` key, so saying a phrase saves nothing through the hook. The code captures the *nearby* text, which may include more than the phrase, is capped (auto-extracts never outrank explicit saves) and runs through the same credential/injection

Context Leakage

High
Category
Data Exfiltration
Confidence
75% confidence
Finding

Code or instructions that leak agent conversation context to external services, potentially exposing sensitive user interactions.

Content

Scanner excerpt · SKILL.md (reported line 181)May include surrounding context.

md
and audit run entirely on your machine — no account and no telemetry by default. Outbound requests can still happen by default — the embedding-model download, the dashboard's automatic version lookup, on hosts with the OpenClaw hook enabled, its npx package resolution, and the agent-config watch's package check for MCP servers it reports as added or changed (see **Outbound connections that happen by default** above); none of them carries memory or conversation content. Because the tool can auto-capture conversation content, here is exactly what it reads, stores, and (only if you opt in) transmits.

- **What it reads.** With the lifecycle handlers enabled (opt-in at setup), ShieldCortex reads your agent **session transcripts — both your prompts and the assistant's replies** — to auto-extract memorable context. PreCompact (before context compaction) reads the recent transcript; the SessionEnd and Stop handlers are **off by default**; the OpenClaw integration extracts from assistant outpu

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 198)May include surrounding context.

md
## What it does NOT do

- Does **not** read SSH keys, AWS credentials, GPG keys, or /etc/ files
- Does **not** send your data to external servers, with three narrow, user-initiated exceptions: Cloud sync when you enable it, the subscription id sent to the licence endpoint when you activate a licence key, and event payloads to webhooks you configure (memory-event and Iron Dome operator notifications). Its remaining outbound calls are downloads/fetches that carry only the request itself — never memory or conversation content: npm-registry update checks, X-Ray package lookups and the OpenClaw hook's npx package resolution (the package name appears in the URL), package-check lookups to registry.npmjs.org, pypi.org and api.osv.dev (the package name and version only), the huggingface.co embedding-model download, and URLs you explicitly pass to `env scan`
- Does **not** modify .bashrc, .zshrc, .profile, or shell configs
- Does **not** use `eval` or dynamic code execution of any kind

Self-Modification

High
Category
Rogue Agent
Confidence
90% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · SKILL.md (reported line 282)May include surrounding context.

Maintenance

bash
shieldcortex update              # Self-update (npm package + OpenClaw plugin + skill)
shieldcortex maintain --dry-run  # Preview storage/memory hygiene (plain `maintain` applies it)

MCP Config Access

High
Category
Agent Snooping
Confidence
95% confidence
Finding

Skill accesses MCP server configuration files (mcp.json). MCP configs contain server URLs, authentication tokens, and tool definitions — reading them allows the skill to discover and potentially abuse other tool integrations.

Content

Scanner excerpt · SKILL.md (reported line 298)May include surrounding context.

md
### `audit` checks:
- **Memory files** — ~/.claude/projects/, ~/.cursor/memories/, ~/.windsurf/memories/
- **Environment** — .env files for leaked credentials (read-only check, never writes)
- **MCP configs** — ~/.claude/mcp.json, ~/.openclaw/mcp.json, ~/.cursor/mcp.json, project-level equivalents
- **Rules files** — CLAUDE.md, .cursorrules, copilot-instructions.md for injection patterns

## What Gets Uploaded to Cloud

Manifest references MCP server without version pin: 'npx -y shieldcortex'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

The manifest includes the unattended, unpinned shieldcortex fallback described in the body. Compromise of a resolved release could execute code with the agent service account's privileges; this duplicates the command-level findings.

Content

No source excerpt is available for this finding.

Manifest references MCP server without version pin: 'npx -y shieldcortex'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

This is a duplicate manifest reference to the unpinned shieldcortex fallback. Automatic acquisition of executable code expands trust beyond the installed release and could enable arbitrary code execution following an upstream compromise.

Content

No source excerpt is available for this finding.

Manifest references MCP server without version pin: 'npx mcporter'.

Medium
Category
MCP Rug Pull
Confidence
97% confidence
Finding

The manifest describes an unpinned mcporter invocation that can fetch and execute code automatically. This is a manifest-level duplicate of the documented hook supply-chain risk, not a separate vulnerability.

Content

No source excerpt is available for this finding.

Manifest references MCP server without version pin: 'npx mcporter'.

Medium
Category
MCP Rug Pull
Confidence
97% confidence
Finding

This repeats the manifest's reference to automatic execution of unpinned mcporter code. The underlying risk is genuine but conditional on remote resolution and upstream compromise, rather than evidence that the current package is malicious.

Content

No source excerpt is available for this finding.

Manifest references MCP server without version pin: 'npx package'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

MCP server references in the skill manifest without version pinning are a rug-pull risk.

Content

No source excerpt is available for this finding.

Manifest references MCP server without version pin: 'npx package'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

MCP server references in the skill manifest without version pinning are a rug-pull risk.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx -y shieldcortex'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

The fallback 'npx -y shieldcortex' permits unattended retrieval and execution of an unpinned package during hook operation. A compromised release could execute under the gateway account without a new approval, independently of the version originally reviewed.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx mcporter'.

Medium
Category
MCP Rug Pull
Confidence
97% confidence
Finding

The documented hook can automatically download and execute an unpinned mcporter release when the package is unavailable locally. A compromised upstream release could consequently run with the gateway user's access to transcripts, configuration, and credentials; disclosure of this behavior does not constrain execution.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx package'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx package'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx package'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx package'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx package'.

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.