Skills backlink-guard Audit History
📦

Audit History

backlink-guard - 4 audits

Version comparison

Capability and finding changes across audited versions, newest first.

VersionDateResultReview itemsChange vs previous
v4 LatestOct 31, 2026, 09:01 AM No confirmed findings1No capability change
v3 Oct 28, 2026, 11:28 AM No confirmed findings0No capability change
v2 Oct 27, 2026, 09:11 AM 1 confirmed0No capability change
v1 Oct 26, 2026, 10:22 AM No confirmed findings1Baseline

Oct 31, 2026, 09:01 AM

Most static alerts match Markdown formatting, documented optional credentials, ordinary local state, and sample URLs rather than malicious behavior. The setup genuinely installs and executes an external PyPI package, whose implementation is absent from the reviewed file. No evidence found of credential theft, obfuscation, or prompt injection; the instructions require approval for paid runs and submissions.

1
Files scanned
190
Lines analyzed
5
Review items
0
False positives ignored
Capability review items (1)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

Medium
Ruby/shell backtick execution
1. Run the pinned release from PyPI: `uvx backlink-guard==0.11.2 <command>` when `uv` is there, else
Setup downloads and executes backlink-guard from PyPI through uvx or pip. Pinning limits version drift but does not prevent compromised code accessing local files and credentials.
Audited by: codex

Oct 27, 2026, 09:11 AM

The static matches are Markdown formatting, ordinary documentation links, application state, credential names, or unrelated prose. The skill instructs users to execute a pinned external package that was not included in this review.

1
Files scanned
189
Lines analyzed
5
Review items
0
False positives ignored

Confirmed security concerns (1)

Medium
Unreviewed external package execution
The instructions download and execute a pinned PyPI package. Its implementation is outside the reviewed skill files, so its behavior and integrity remain unverified.
The instructions explicitly invoke a PyPI release, and the reviewed file set contains only the skill documentation.
Audited by: codex

Oct 26, 2026, 10:22 AM

Of 105 static findings, 104 are false positives involving Markdown, optional credential names, documented URLs, application state, or SEO terminology. One finding identifies the external-package execution boundary: the required PyPI runtime is absent from the audited files, so its behavior remains unverified. No evidence found of prompt injection, credential theft, destructive commands, or unauthorized reporting in SKILL.md.

1
Files scanned
177
Lines analyzed
5
Review items
0
False positives ignored
Capability review items (1)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

Medium
Ruby/shell backtick execution
1. Run the pinned release from PyPI: `uvx backlink-guard==0.10.0 <command>` when `uv` is there, else
The setup explicitly installs or executes a PyPI package absent from the audited files. Pinning does not verify its behavior with local access or optional credentials.
Audited by: codex