The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external research agents to generate new documentation and code templates, creating a vulnerability to malicious instructions embedded in that data.
Ingestion points: As described in SKILL.md and references/workflow.md, the skill mandates the use of research-expert to fetch official documentation from the internet and explore-codebase to analyze local skill structures.
Boundary markers: The instructions lack specific requirements for using delimiters or boundary markers to separate untrusted research data from the agent's core generation logic.
Capability inventory: The skill possesses the capability to create directories (mkdir), copy files (cp), delete files (rm), and modify file content using stream editors (sed) as part of its scaffolding workflow.
Sanitization: There are no documented procedures for sanitizing, escaping, or validating the external content gathered during the research phase before it is used to populate new skill files.
[COMMAND_EXECUTION]: The skill provides instructions and examples for executing shell-based file system and text manipulation commands.
Evidence: references/adaptation.md demonstrates the use of cp -r, rm, and sed -i to automate the adaptation of skills between different frameworks, such as converting Next.js patterns to React.

skill-creator