The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (LOW): The skill performs an installation of markdownlint-cli2 via npm install -g. While this downloads code from a remote registry, the package is a well-known industry standard for markdown linting. Following the [TRUST-SCOPE-RULE], this is classified as LOW severity.
[COMMAND_EXECUTION] (SAFE): The skill uses Bash to execute markdownlint-cli2 commands. These actions are directly tied to the primary purpose of fixing documentation and are not used for malicious purposes.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill ingests untrusted markdown data from the project.
Ingestion points: Reads *.md files via Glob and Read tools.
Boundary markers: None explicitly defined in the prompts to separate file content from instructions.
Capability inventory: Includes Bash, Write, and Edit which could be misused if linter output was successfully poisoned.
Sanitization: Not present; the skill relies on the linter's structured output but processes content through automated fixes.

markdown-linter-fixer