The Agent Skills Directory

[SAFE]: The skill is a developer tool for validating 'SKILL.md' files. It uses standard tools (Read, Glob, Grep) to perform static analysis on file content.
[PROMPT_INJECTION]: While the skill contains documentation about prompt injection (e.g., in references/marketplace-checks.md and SKILL.md under the OWASP section), these are instructional examples and detection rules meant to find injections in other skills, not an attempt to inject the current agent context.
[EXTERNAL_DOWNLOADS]: The references/cli.md file references a Go CLI tool hosted on GitHub (github.com/olgasafonova/skillcheck-cli). This is a legitimate project resource belonging to the skill's author and is documented as a standalone installation option, not a silent background download.
[COMMAND_EXECUTION]: The skill documentation includes example shell commands (e.g., go install, skillcheck fix) and references dangerous commands like rm -rf or DROP TABLE within its 'OWASP Agentic Top 10' validation logic. These are used for pattern matching and educational purposes to warn users about risks in their own skills; they are not executed by the skill itself.
[DATA_EXFILTRATION]: There are no network-enabled tools or instructions to send data externally. The analysis is performed locally/offline.
[NO_CODE]: The skill consists entirely of Markdown instructions and documentation. It does not ship with executable scripts or binaries.

skill-check