📦

Audit History

dependency-upgrade - 9 audits

Version comparison

Capability and finding changes across audited versions, newest first.

VersionDateResultReview itemsChange vs previous
v9 LatestJul 7, 2026, 07:23 AM 1 confirmed5No capability change
v8 Jul 7, 2026, 07:23 AM 1 confirmed5No capability change
v7 Jun 30, 2026, 10:50 PM No confirmed findings4No capability change
v6 Jun 30, 2026, 10:50 PM No confirmed findings4External commandsNetwork accessFilesystem access
v5 Jan 21, 2026, 07:15 PM 1 confirmed0 Network accessExternal commandsFilesystem access
v4 Jan 17, 2026, 08:22 AM No confirmed findings0No capability change
v3 Jan 17, 2026, 08:22 AM No confirmed findings0Network accessExternal commandsFilesystem access
v2 Jan 4, 2026, 05:06 PM No confirmed findings0No capability change
v1 Jan 4, 2026, 05:06 PM No confirmed findings0Baseline

Jul 7, 2026, 07:23 AM

Most static external-command findings are false positives caused by Markdown code fences, not actual Ruby or shell backtick execution. No evidence found for prompt injection, credential exfiltration, or hidden malicious instructions. The skill includes legitimate high-impact package, network, and filesystem examples, so users should review commands and diffs before execution.

1
Files scanned
410
Lines analyzed
9
Review items
0
False positives ignored

Confirmed security concerns (1)

Medium
High-Impact Upgrade Commands Require Guardrails
SKILL.md includes commands that update packages, run remote package executables, use force install flags, and delete an upgrade branch. These are legitimate upgrade actions but can modify project state or hide dependency conflicts if run without review.
The cited lines contain explicit npm, npx, git, and force-install commands. The guidance is relevant to dependency upgrades, so the risk is operational rather than malicious.
Capability review items (5)

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
Node.js fs operations
let content = fs.readFileSync(file, 'utf8');
The migration script example reads project files matched by a glob. It is legitimate migration guidance, but it can access local source files if users run the example.
Medium
Node.js fs operations
fs.writeFileSync(file, content);
The migration script example writes modified content back to source files. This is expected for codemods, but it can overwrite project files without a dry run or review.
Medium
Synchronous file operations
let content = fs.readFileSync(file, 'utf8');
The migration script example reads project files matched by a glob. It is legitimate migration guidance, but it can access local source files if users run the example.
Medium
Synchronous file operations
fs.writeFileSync(file, content);
The migration script example writes modified content back to source files. This is expected for codemods, but it can overwrite project files without a dry run or review.
Low
Hardcoded URL
curl https://raw.githubusercontent.com/facebook/react/main/CHANGELOG.md
The skill shows a curl command to fetch a React changelog from a hardcoded GitHub URL. The target is public documentation, but executing it still requires external network access and review.
Audited by: codex

Jul 7, 2026, 07:23 AM

Most static external-command findings are false positives caused by Markdown code fences, not actual Ruby or shell backtick execution. No evidence found for prompt injection, credential exfiltration, or hidden malicious instructions. The skill includes legitimate high-impact package, network, and filesystem examples, so users should review commands and diffs before execution.

1
Files scanned
410
Lines analyzed
9
Review items
0
False positives ignored

Confirmed security concerns (1)

Medium
High-Impact Upgrade Commands Require Guardrails
SKILL.md includes commands that update packages, run remote package executables, use force install flags, and delete an upgrade branch. These are legitimate upgrade actions but can modify project state or hide dependency conflicts if run without review.
The cited lines contain explicit npm, npx, git, and force-install commands. The guidance is relevant to dependency upgrades, so the risk is operational rather than malicious.
Capability review items (5)

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
Node.js fs operations
let content = fs.readFileSync(file, 'utf8');
The migration script example reads project files matched by a glob. It is legitimate migration guidance, but it can access local source files if users run the example.
Medium
Node.js fs operations
fs.writeFileSync(file, content);
The migration script example writes modified content back to source files. This is expected for codemods, but it can overwrite project files without a dry run or review.
Medium
Synchronous file operations
let content = fs.readFileSync(file, 'utf8');
The migration script example reads project files matched by a glob. It is legitimate migration guidance, but it can access local source files if users run the example.
Medium
Synchronous file operations
fs.writeFileSync(file, content);
The migration script example writes modified content back to source files. This is expected for codemods, but it can overwrite project files without a dry run or review.
Low
Hardcoded URL
curl https://raw.githubusercontent.com/facebook/react/main/CHANGELOG.md
The skill shows a curl command to fetch a React changelog from a hardcoded GitHub URL. The target is public documentation, but executing it still requires external network access and review.
Audited by: codex

Jun 30, 2026, 10:50 PM

Static analysis flagged many shell commands, one network URL, file operations, and weak cryptography patterns in SKILL.md. Review found these are instructional examples for dependency upgrades, with weak cryptography alerts caused by ordinary words such as description or describe. The skill is publishable with a medium warning because copied commands can install packages, rewrite files, fetch remote changelogs, or remove local branches.

1
Files scanned
410
Lines analyzed
7
Review items
1
False positives ignored
Capability review items (4)

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
Package Manager Commands Modify Dependencies
The skill shows npm, yarn, npx, and install commands that can modify dependency manifests, lock files, or installed packages. This is expected for the skill purpose, but users should review commands before execution and avoid untrusted package names.
The commands are visible in documentation blocks and directly change package state. The surrounding context is legitimate dependency upgrade guidance, so this is operational risk rather than malicious intent.
Medium
Remote Content Retrieval Example
The skill includes a curl command for reading a public React changelog. The URL is relevant to upgrade research, but remote content should be inspected before use in automation.
There is direct evidence of a network command to GitHub. It fetches public documentation and does not transmit secrets, so the risk is moderate.
Medium
File Rewriting Migration Script Example
The migration script example reads and writes matching source files while replacing React API usage. This is normal migration guidance, but broad globs can unintentionally rewrite many files.
The fs.readFileSync and fs.writeFileSync calls are explicit and operate over src/**/*.tsx. The code is illustrative, but copied use could alter project files.
Medium
Rollback Script Contains Destructive Git Commands
The rollback example uses git stash, branch checkout, and branch deletion. These commands support rollback planning, but users should verify branch names and uncommitted work first.
The shell example clearly changes repository state and can delete a branch. The commands are not hidden and match the documented rollback scenario.
Static false positives ignored (1)

These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.

Low
Weak Cryptography Alerts Are False Positives
Static analysis reported weak cryptographic algorithm indicators at several lines, but the reviewed lines contain prose, front matter, test names, or package upgrade examples. No cryptographic API usage or insecure hashing implementation was found.
Manual review of the cited lines found no crypto functions or algorithm configuration. The scanner appears to match substrings inside ordinary text such as description or describe.

Detected Patterns

Unpinned Latest Package InstallationRecursive Source File Writes
Audited by: codex

Jun 30, 2026, 10:50 PM

Static analysis flagged many shell commands, one network URL, file operations, and weak cryptography patterns in SKILL.md. Review found these are instructional examples for dependency upgrades, with weak cryptography alerts caused by ordinary words such as description or describe. The skill is publishable with a medium warning because copied commands can install packages, rewrite files, fetch remote changelogs, or remove local branches.

1
Files scanned
410
Lines analyzed
7
Review items
1
False positives ignored
Capability review items (4)

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
Package Manager Commands Modify Dependencies
The skill shows npm, yarn, npx, and install commands that can modify dependency manifests, lock files, or installed packages. This is expected for the skill purpose, but users should review commands before execution and avoid untrusted package names.
The commands are visible in documentation blocks and directly change package state. The surrounding context is legitimate dependency upgrade guidance, so this is operational risk rather than malicious intent.
Medium
Remote Content Retrieval Example
The skill includes a curl command for reading a public React changelog. The URL is relevant to upgrade research, but remote content should be inspected before use in automation.
There is direct evidence of a network command to GitHub. It fetches public documentation and does not transmit secrets, so the risk is moderate.
Medium
File Rewriting Migration Script Example
The migration script example reads and writes matching source files while replacing React API usage. This is normal migration guidance, but broad globs can unintentionally rewrite many files.
The fs.readFileSync and fs.writeFileSync calls are explicit and operate over src/**/*.tsx. The code is illustrative, but copied use could alter project files.
Medium
Rollback Script Contains Destructive Git Commands
The rollback example uses git stash, branch checkout, and branch deletion. These commands support rollback planning, but users should verify branch names and uncommitted work first.
The shell example clearly changes repository state and can delete a branch. The commands are not hidden and match the documented rollback scenario.
Static false positives ignored (1)

These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.

Low
Weak Cryptography Alerts Are False Positives
Static analysis reported weak cryptographic algorithm indicators at several lines, but the reviewed lines contain prose, front matter, test names, or package upgrade examples. No cryptographic API usage or insecure hashing implementation was found.
Manual review of the cited lines found no crypto functions or algorithm configuration. The scanner appears to match substrings inside ordinary text such as description or describe.

Detected Patterns

Unpinned Latest Package InstallationRecursive Source File Writes
Audited by: codex

Jan 21, 2026, 07:15 PM

This skill provides documentation and guidance for dependency management workflows. All static findings are false positives: the external commands are example bash/npm commands in documentation, filesystem operations are example migration scripts, and weak crypto detections are from the word 'MD5' in version strings. The skill contains no executable code and poses no security risk.

2
Files scanned
1,190
Lines analyzed
1
Review items
0
False positives ignored

Confirmed security concerns (1)

Low
Documentation Contains Command Examples
The SKILL.md file contains numerous example bash commands for dependency management (npm, yarn, git). These are educational examples showing proper dependency upgrade workflows, not executable code. Users must manually execute these commands in their own environments.
Audited by: claude

Jan 4, 2026, 05:06 PM

Pure documentation skill providing guidance on dependency upgrades. Contains no executable code, network calls, file access, or command execution. All commands shown are illustrative examples for user reference.

4
Files scanned
605
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were recorded for this completed audit.
Audited by: claude

Jan 4, 2026, 05:06 PM

Pure documentation skill providing guidance on dependency upgrades. Contains no executable code, network calls, file access, or command execution. All commands shown are illustrative examples for user reference.

4
Files scanned
605
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were recorded for this completed audit.
Audited by: claude