The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a shell script involving find, grep, sed, and tr to process local files and generate documentation. It also invokes markdownlint-cli via bunx to validate the resulting README.md file.- [INDIRECT_PROMPT_INJECTION]: This skill has a surface for indirect prompt injection as it aggregates content from multiple files into a single document. * Ingestion points: Metadata (descriptions) are read from all SKILL.md files within the skills/ subdirectory. * Boundary markers: The skill does not use any delimiters or warnings to isolate the ingested content within the generated table. * Capability inventory: The skill can read arbitrary files within the specified directory structure and output formatted markdown. * Sanitization: The script performs minimal sanitization (removing quotes and truncating length), which does not protect against malicious markdown or prompt injection payloads embedded in the source descriptions.

readme-sync