Skylv Dependency Updater

Auto-checks and updates outdated dependencies. Shows changelogs and breaking changes before updating. Triggers: update dependencies, upgrade packages, check...

Security Scan

The skill's purpose (update dependencies) matches its instructions, but the runtime instructions reference filesystem access, command-line tools, test execution, and repo commits without declaring required binaries, credentials, or safeguards — this mismatch and the potential to run arbitrary project code warrant caution.

Purpose & Capability

Name/description match the instructions: it scans for package manifests and runs package-manager checks. However, the metadata lists no required binaries or config paths while the SKILL.md expects npm, pip, cargo, go (and implicitly git and test runners) to be available. The lack of declared runtime dependencies is a documentation/incoherence issue.

Instruction Scope

Instructions direct the agent to detect package managers by reading project files (package.json, pyproject.toml, etc.), run commands that may contact registries (npm outdated, pip list --outdated, cargo outdated), run tests after each update, and commit changes. Running tests and committing changes executes code and writes to the repository — reasonable for an updater but high-privilege. The SKILL.md doesn't require explicit user confirmation, sandboxing, or safety checks before executing these actions.

Install Mechanism

Instruction-only skill with no install spec, so nothing is written to disk by the skill itself. This is lowest install risk. However, some referenced tools (e.g., cargo-outdated) may be third-party subcommands not present by default; the skill doesn't explain how to obtain them.

Credentials

The skill requests no environment variables or credentials, which is consistent at face value. In practice, updating packages in projects that use private registries may require auth tokens (npm, pip indexes) or git credentials to push commits; the SKILL.md does not mention or justify access to such credentials or how they will be used.

Persistence & Privilege

The skill is not always-enabled and has no install-time persistence. It does not request system-wide configuration changes or access to other skills' configs.

What to consider before installing

This skill appears to do what it says, but it has gaps and potential risks you should consider before enabling it: - The SKILL.md instructs the agent to read project files, run package-manager commands, run tests, and make git commits. Those are reasonable for a dependency updater but are high-privilege operations because tests and scripts in a repo can execute arbitrary code. - The skill metadata did not declare required binaries (npm, pip, cargo, go, git, test runners) or mention how it handles private registries/auth tokens. Expect runtime failures or unexpected behavior if those tools/credentials are missing. - Ask or require the agent to prompt for explicit confirmation before making any updates, running tests, or committing changes. Prefer running updates in a disposable/sandboxed environment and creating backups or branches before changes. - If your projects use private package registries or require git push access, consider restricting which credentials the skill can use and only provide scoped tokens if you decide to proceed. - If you are not comfortable with the agent running tests or committing changes autonomously, do not enable autonomous invocation for this skill and require manual review of its report before any update actions. If the publisher can (a) declare required binaries, (b) describe credential usage and confirmation prompts, and (c) document sandboxing/testing safeguards, the risk would be much lower.

Dependency Updater

Overview

Scans project dependencies and checks for updates, shows changelogs, identifies breaking changes.

When to Use

• User asks to "update dependencies" or "check for updates"
• Regular maintenance

How It Works

Step 1: Detect package manager

package.json -> npm pyproject.toml -> pip Cargo.toml -> cargo go.mod -> go

Step 2: Check outdated

npm: npm outdated --json pip: pip list --outdated --format=json cargo: cargo outdated

Step 3: Risk assessment

Patch (1.2.3 -> 1.2.4): Low risk - auto-update Minor (1.2.3 -> 1.3.0): Medium - show changelog Major (1.2.3 -> 2.0.0): High - show breaking changes

Output Format

Major Updates: express 4.17.1 -> 5.0.0 [BREAKING changes] Minor Updates: axios 0.21.1 -> 0.21.4 [Bug fixes] Patch Updates: debug 4.3.1 -> 4.3.4 [Security patch]

Update Strategy

1. Show report first - never update blindly
2. Update in stages: patches -> minors -> majors
3. Run tests after each update
4. Commit each update separately