The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and validate untrusted Markdown files which could contain malicious instructions intended to influence the agent's logic during the audit or repair process.
Ingestion points: The skill reads file content via gbrain frontmatter validate and gbrain frontmatter audit commands.
Boundary markers: No specific boundary markers or instructions to ignore embedded commands within the scanned files are mentioned.
Capability inventory: The skill uses the exec tool to run CLI commands and can write modifications to the filesystem using the --fix flag.
Sanitization: The skill relies on the external gbrain CLI tool for YAML parsing and validation, but the agent itself interprets the tool's output.
[PERSISTENCE]: The gbrain frontmatter install-hook command installs a git pre-commit hook in the user's repository. This ensures that the validation logic runs automatically during git operations, creating a persistent execution point within the developer's workflow.
[COMMAND_EXECUTION]: The skill uses the exec tool to invoke the gbrain CLI. While these commands are scoped to frontmatter validation and repair, they involve executing shell commands with arguments derived from file paths and source identifiers.

frontmatter-guard