Security audit
Security checks for vulnerabilities and agentic risk
The skill is a coherent Discogs account-management tool, but it deserves review because it can change account data, stores a long-lived token in local plaintext config, and pins dependencies that scanner evidence flags as vulnerable.
Before installing, prefer using DISCOGS_USER_TOKEN instead of storing the token with auth, revoke the token if the machine is shared or compromised, update or audit the pinned dependencies, and use --dry-run or preview output before bulk sync or any removal.
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
## Setup
### 1. Generate a Personal Access Token
1. Go to https://www.discogs.com/settings/developers
2. Click "Generate new token"
The description presents a broader sync/manage skill for Discogs wantlists and collections, with marketplace pricing as one capability among several. The supplied code chunk, however, is narrowly focused on marketplace statistics retrieval and related release/version resolution. It does not perform any write operations to a wantlist or collection, nor list those resources. This is a material description-behavior mismatch for this chunk because the declared primary functionality emphasizes collection/wantlist management, while the code here only supports marketplace search/pricing. The code does remain within Discogs-related API behavior and is consistent with the marketplace-pricing portion of the description, so the mismatch is due to missing declared capabilities rather than suspicious extra behavior.
The dependency list pins click==8.3.1, and the supplied advisory indicates versions 8.3.2 and below are affected by a command injection vulnerability. In a CLI-driven skill that processes user-provided arguments and potentially bulk file input, a vulnerable command-line framework can increase the risk that crafted input triggers unsafe command handling paths.
The requirements pin idna==3.11, which the advisory reports as affected by specially crafted IDNA input handling flaws. This skill is intended to contact only api.discogs.com, so exposure is narrower than in a general-purpose URL processor, but malformed internationalized host/domain handling can still matter anywhere HTTP libraries resolve or validate hostnames.
oauthlib==3.3.1 is flagged with multiple advisories including JSONP callback injection and PKCE timing issues. The skill description references use of a Discogs personal access token rather than a full OAuth browser flow, which reduces practical exposure, but shipping a known-vulnerable auth library is still a supply-chain and future-maintenance risk if any OAuth-related paths are present.
urllib3==2.6.3 is associated with multiple high-severity advisories, including TLS proxy handling issues, infinite-loop conditions, and decompression-related weaknesses. Because this skill is network-facing and communicates with an external API over HTTPS, weaknesses in the HTTP transport layer materially increase risk of denial of service, traffic handling failures, or weakened connection guarantees.
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.
"""Authentication flow for Discogs (personal access token)."""
from __future__ import annotations
This code persists a long-lived personal access token in a plaintext JSON config file on disk. Although the file permissions are tightened on non-Windows systems, plaintext secret storage increases the risk of credential exposure through local compromise, backups, sync tools, or weaker permission handling on other platforms.
def save_user_token(user_token: str, username: str | None = None) -> None:
"""Store a personal access token to config file."""
config = load_config()
config.update(
{
Without declared permissions the skill's intent is opaque and cannot be validated.
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.
click==8.3.1 — CLI frameworkrich==14.3.3 — Terminal output formattingInstallation (one-time, done by the user — not by the agent): create a venv at ~/.discogs-sync/venv and install the pinned requirements. discogs-sync.py picks up this venv automatically; it lives outside the skill directory so skill updates don't remove it.
python3 -m venv ~/.discogs-sync/venv
Pygments==2.19.2 is reported vulnerable to a Regular Expression Denial of Service. In this skill, Pygments is likely only an indirect dependency used for terminal formatting, so the context makes exploitation less likely unless untrusted text is highlighted or parsed through vulnerable lexers.
requests==2.32.5 is flagged for insecure temporary file reuse in a utility function related to zipped paths. The skill's normal Discogs API usage may not invoke that specific helper, which lowers immediate exploitability, but retaining a dependency version with a known flaw is still unsafe, especially if future code adds archive handling for bulk inputs or downloaded content.
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.
class ConfirmationRequiredError(SyncError):
"""A removal was requested without confirmation; nothing was changed.
``preview`` describes exactly what would have been removed.
"""
The function performs a write to the user's Discogs collection by calling _add_to_collection, but this path has no confirmation prompt, print/log disclosure, or warning in the local docstring about modifying remote user data. Because this is a state-changing remote operation, the lack of explicit user disclosure in this code path meets the missing-warning criteria for code files.
No suspicious patterns detected.