Back to skill

Security audit

Musicbrainz Importer

Security checks for vulnerabilities and agentic risk

Overview

The skill does what it claims for MusicBrainz, but it asks users to store account passwords in plaintext and uses unpinned browser automation for account-changing actions.

Review before installing. Use a dedicated MusicBrainz password that is not reused elsewhere, prefer environment variables or a secure secret store over .credentials.json, restrict any credential file to owner-only permissions, and verify every artist, release, and cover-art edit before the skill submits it. Pin Playwright or install from a locked, trusted project directory if possible.

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 (3)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:14
Finding

Unpinned Playwright Dependency Installation

Content
View full analysis
/dev/null || npm install playwright ``` ### Technical Analysis The Skill installs `playwright` without an exact version or a lockfile. Consequently, the code downloaded and executed can change after the Skill has been reviewed. npm installation may also run package lifecycle scripts with the privileges of the Agent process. Installation takes place from the shared `/tmp` directory rather than a dedicated project directory. This makes package resolution and generated npm state less controlled and may permit interference by other local users or processes on a multi-user system. No evidence indicates that the current Playwright package is malicious. The vulnerability is the mutable, unverified dependency-installation process. ### Attack Path 1. An attacker compromises the upstream package, an associated dependency, or the relevant registry delivery path. 2. A user or Agent follows the setup instructions and executes `npm install playwright`. 3. npm resolves the latest package version because no exact version or lockfile is specified. 4. The compromised package or dependency is downloaded. 5. Malicious installation or lifecycle code executes with the permissions of the Agent account. 6. That code can access files, credentials, and network resources available to the Agent. A local attacker may also attempt to manipulate npm-related state under the shared temporary directory before installation, depending on directory permissions and the surrounding environment. ### Impact Assessment Successful exploitation could provi ...[truncated 377 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:76
Finding

Plaintext Credential Storage Without Permission Enforcement

Content
View full analysis
~/.openclaw/skills/musicbrainz/.credentials.json << 'EOF' {"username": "USER", "password": "PASS"} EOF ``` From `README.md`: ```bash cat > ~/.hermes/skills/openclaw-imports/musicbrainz-importer/.credentials.json << 'EOF' {"username": "YOUR_MB_USERNAME", "password": "YOUR_MB_PASSWORD"} EOF ``` From `scripts/preflight.sh`: ```bash SKILL_DIR="$(cd "$(dirname "$0")/.." && pwd)" CRED_FILE="$SKILL_DIR/.credentials.json" # Load from credentials file if it exists if [ -f "$CRED_FILE" ]; then FILE_USER=$(jq -r '.username // empty' "$CRED_FILE" 2>/dev/null || true) FILE_PASS=$(jq -r '.password // empty' "$CRED_FILE" 2>/dev/null || true) if [ -n "$FILE_USER" ] && [ -n "$FILE_PASS" ]; then export MB_USERNAME="$FILE_USER" export MB_PASSWORD="$FILE_PASS" fi fi ``` ### Technical Analysis The documentation instructs users to persist their MusicBrainz username and password as plaintext JSON. Neither the creation instructions nor `preflight.sh` set or validate owner-only file permissions. The resulting permissions therefore depend on the user's current `umask` and the permissions of the containing directory. On a multi-user system, a permissively created file may be readable by other users or processes. Plaintext files may also be collected by backup, indexing, synchronization, or diagnostic software. The credential transmission detected by the static pre-scan is not, by itself, malicious. `scripts/preflight.sh:35-43` sends the username and password over HTTPS to the declared official MusicBrainz login endpoint for authentication validation. That network behavior is consistent with the Skill's write functionality. The confirmed weakness is local storage and permission handling, not transmi ...[truncated 927 chars]
Remediation
View remediation
~/.openclaw/skills/musicbrainz/.credentials.json <<'EOF' {"username": "USER", "password": "PASS"} EOF chmod 600 ~/.openclaw/skills/musicbrainz/.credentials.json ``` 3. Update `preflight.sh` to verify that the credential file is owned by the current user and is not readable or writable by group or others. 4. Refuse to load credentials when ownership or permissions are unsafe. 5. Avoid exporting the password longer than necessary; keep it scoped to the authentication operation. 6. Clearly warn users not to reuse the MusicBrainz password on other services. 7. Ensure the credentials file is excluded from source control, backups, logs, and artifact packaging where appropriate. ]]>

T09 · Insecure Skill Coding Practices

Note
Location
scripts/seed_release.mjs:142
Finding

Predictable Temporary Output Permits Symlink and File-Overwrite Attacks

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (22)

Credential Access

High
Category
Privilege Escalation
Confidence
98% confidence
Finding

The README explicitly shows storing a username and password in .credentials.json, creating a direct credential exposure risk. Because this skill performs write operations against a web account and may use browser automation, compromise of that file could enable unauthorized edits, account takeover of the MusicBrainz account session, or reuse attacks if the password is shared across services.

Content

Scanner excerpt · README.md (reported line 47)May include surrounding context.

Store your MusicBrainz credentials for write operations:

bash
cat > ~/.hermes/skills/openclaw-imports/musicbrainz-importer/.credentials.json << 'EOF'
{"username": "YOUR_MB_USERNAME", "password": "YOUR_MB_PASSWORD"}
EOF

Credential Access

High
Category
Privilege Escalation
Confidence
97% confidence
Finding

This instruction establishes a local credential file as part of the workflow, meaning the skill is explicitly designed to access sensitive authentication material. In the context of shell execution and browser automation, local credential access materially increases the blast radius of misuse or compromise.

Content

Scanner excerpt · SKILL.md (reported line 76)May include surrounding context.

bash scripts/preflight.sh

text

Credentials live in `.credentials.json` in the skill directory.

- **`MISSING_CREDENTIALS`**: Ask user for MusicBrainz username + password. Save:
  ```bash

Credential Access

High
Category
Privilege Escalation
Confidence
99% confidence
Finding

The cat > ~/.openclaw/skills/musicbrainz/.credentials.json example writes a username and password directly to disk in plaintext. This is dangerous because secrets may persist in shell history, backups, endpoint telemetry, or be readable by other local processes if permissions are weak.

Content

Scanner excerpt · SKILL.md (reported line 80)May include surrounding context.

  • MISSING_CREDENTIALS: Ask user for MusicBrainz username + password. Save:
    bash
    cat > ~/.openclaw/skills/musicbrainz/.credentials.json << 'EOF'
    {"username": "USER", "password": "PASS"}
    EOF
    

Credential Access

High
Category
Privilege Escalation
Confidence
95% confidence
Finding

The Playwright login script reads credentials from a predictable local file path and uses them for authenticated automation. Predictable secret locations are high-risk because any other skill or process with filesystem access can target the same path and exfiltrate or misuse the account.

Content

Scanner excerpt · SKILL.md (reported line 96)May include surrounding context.

javascript
import { readFileSync } from 'fs';
const creds = JSON.parse(readFileSync(
  process.env.HOME + '/.openclaw/skills/musicbrainz/.credentials.json'
));

await page.goto('https://musicbrainz.org/login', { waitUntil: 'networkidle', timeout: 30000 });

Credential Access

High
Category
Privilege Escalation
Confidence
87% confidence
Finding

Even though this line only reads the username from the credential file for a User-Agent string, it normalizes access to the same secret-bearing file during routine lookup operations. Expanding credential-file reads into read-only paths increases unnecessary exposure and can accidentally leak account-identifying data to external services.

Content

Scanner excerpt · SKILL.md (reported line 513)May include surrounding context.

Search by name (no Spotify link)

bash
# Build User-Agent: use bot name + MB username from .credentials.json if available
MB_USER=$(jq -r '.username // empty' ~/.openclaw/skills/musicbrainz/.credentials.json 2>/dev/null || true)
UA="${OPENCLAW_BOT_NAME:-OpenClawBot}/1.0 (${MB_USER:-anonymous}@users.noreply.musicbrainz.org)"

Credential Access

High
Category
Privilege Escalation
Confidence
84% confidence
Finding

Constructing the User-Agent with a username derived from the credentials file sends account-identifying information to a remote service during API access. While less severe than exposing a password, it still unnecessarily couples secret storage with network-visible identity and broadens privacy and targeting risks.

Content

Scanner excerpt · SKILL.md (reported line 514)May include surrounding context.

bash
# Build User-Agent: use bot name + MB username from .credentials.json if available
MB_USER=$(jq -r '.username // empty' ~/.openclaw/skills/musicbrainz/.credentials.json 2>/dev/null || true)
UA="${OPENCLAW_BOT_NAME:-OpenClawBot}/1.0 (${MB_USER:-anonymous}@users.noreply.musicbrainz.org)"

curl -s "https://musicbrainz.org/ws/2/artist/?query=<name>&fmt=json&limit=5" \

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

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

Content

Scanner excerpt · references/api.md (reported line 13)May include surrounding context.

md
# Load MB username from credentials file if not already set
if [ -z "${MB_USERNAME:-}" ]; then
  SKILL_DIR="$(cd "$(dirname "$0")/.." && pwd)"
  MB_USERNAME=$(jq -r '.username // empty' "$SKILL_DIR/.credentials.json" 2>/dev/null || true)
fi
CONTACT="${MB_USERNAME:-anonymous}@users.noreply.musicbrainz.org"
UA="${OPENCLAW_BOT_NAME:-OpenClawBot}/1.0 ($CONTACT)"

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

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

Content

Scanner excerpt · scripts/mb_lookup.sh (reported line 12)May include surrounding context.

sh
# Load MB username from credentials file if not already set
if [ -z "${MB_USERNAME:-}" ]; then
  SKILL_DIR="$(cd "$(dirname "$0")/.." && pwd)"
  MB_USERNAME=$(jq -r '.username // empty' "$SKILL_DIR/.credentials.json" 2>/dev/null || true)
fi
CONTACT="${MB_USERNAME:-anonymous}@users.noreply.musicbrainz.org"
UA="${OPENCLAW_BOT_NAME:-OpenClawBot}/1.0 ($CONTACT)"

Credential Access

High
Category
Privilege Escalation
Confidence
87% confidence
Finding

The script reads a local .credentials.json file and exports the username and password into environment variables, centralizing access to reusable account credentials. This creates a sensitive credential handling path that can be abused if the file is weakly protected, accidentally committed, or if child processes, logs, or diagnostics expose the exported secrets.

Content

Scanner excerpt · scripts/preflight.sh (reported line 8)May include surrounding context.

sh
set -euo pipefail

SKILL_DIR="$(cd "$(dirname "$0")/.." && pwd)"
CRED_FILE="$SKILL_DIR/.credentials.json"

# Load from credentials file if it exists
if [ -f "$CRED_FILE" ]; then

Rp1

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.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · README.md (reported line 18)May include surrounding context.

md
- **Node.js** (for Playwright scripts)
- **Playwright** with Chromium (`npx playwright install chromium`)
- **curl** and **jq** (for API lookups)
- A [MusicBrainz account](https://musicbrainz.org/register) (for write operations)

## Install

Rp1

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.

Rp1

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.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README instructs users to place MusicBrainz credentials in a plaintext local file under the skill directory without warning about file permissions, secret handling, or exclusion from version control. In a skill ecosystem, this is dangerous because other local tools, backups, logs, or accidental repository commits could expose the credentials.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill invokes shell commands extensively (bash, curl, npm, find) but does not declare an explicit tool scope or allowed-tools policy. That mismatch increases the chance an agent will execute powerful local commands without clear sandboxing or least-privilege constraints.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger description includes broad phrases like 'music database' and 'music metadata', which can cause the skill to activate for loosely related requests. Since the skill performs shell execution, credential handling, and write actions to a third-party account, overbroad triggering raises the risk of accidental invocation in contexts the user did not intend.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
88% confidence
Finding

The skill instructs use of npx playwright/npm-installed Playwright without pinning a version, which makes executions non-reproducible and exposes the workflow to supply-chain drift or a compromised upstream release. Because this skill performs authenticated browser automation, a malicious dependency update could affect credentialed sessions.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 60)May include surrounding context.

md
There are **four operations**. Each is a self-contained Playwright script. Never combine them into one script. Run them sequentially:

1. **Login** — authenticate the browser session
2. **Add Artist** — create a new artist entity
3. **Add Release** — create a release with tracks, metadata, and Spotify link
4. **Upload Cover Art** — attach album artwork to an existing release

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill tells the operator to save a MusicBrainz username and password into a local JSON file but does not prominently warn about plaintext secret storage or handling risks. This can lead to credential exposure through local compromise, backups, logs, or overly broad file permissions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The guidance presents fixed values for language (120 = English) and script (28 = Latin), and the sample release-creation code sets those exact defaults. This imposes a specific language/locale assumption without offering opt-in or clarifying that these fields should be chosen based on the actual release metadata.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
94% confidence
Finding

The script transmits MusicBrainz credentials to an external service during a preflight check by performing an actual login request. Even though the destination appears legitimate and uses HTTPS, this still exposes stored credentials over the network outside the user's immediate action flow and increases risk from credential leakage via process inspection, logging, interception by local tooling, or unintended execution in automated environments.

Content

Scanner excerpt · scripts/preflight.sh (reported line 35)May include surrounding context.

sh
trap "rm -f $COOKIE_JAR" EXIT

# Fetch login page to get session cookie
curl -s -c "$COOKIE_JAR" -o /dev/null "https://musicbrainz.org/login"

# Attempt login
RESPONSE=$(curl -s -w "\n%{http_code}" \

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

The markdown tells the operator to use curl to save cover art to /tmp/cover.jpg, which is a local file write, but the skill description does not explicitly call out that temporary files are created. Since file writes can affect user system state, the behavior should be disclosed in the instructions.

Content

No source excerpt is available for this finding.