The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted repository data while having access to execution tools.
Ingestion points: As defined in SKILL.md and references/techniques.md, the agent researches the codebase using tools like ls, grep, and cat (via Read) to understand project structure and analyze existing implementations.
Boundary markers: There are no explicit delimiters or specific instructions for the agent to ignore potentially malicious content or instructions embedded within the files it reads during the research phase.
Capability inventory: The skill is configured in SKILL.md with capabilities including Bash(node:*) (allowing execution of Node.js scripts) and Write (allowing file system modifications).
Sanitization: No sanitization, validation, or escaping of the ingested code content is specified in the workflow prompts before the agent processes the information.

codex-brainstorm