The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill contains templates that instruct the AI to ingest and analyze code provided by external contractors. This introduces a potential surface for indirect prompt injection where malicious instructions hidden in the contractor's code (e.g., in comments or strings) could attempt to manipulate the agent's analysis.
Ingestion points: The prompt template in SKILL.md under the 'The Paid Trial' section uses placeholders for the user to [paste code or point to files] for AI review.
Boundary markers: The provided prompt templates do not include specific delimiters or 'ignore instructions' warnings to isolate the untrusted code from the agent's operational instructions.
Capability inventory: The skill has no defined capabilities or tool access; it consists entirely of instructional markdown and prompt templates.
Sanitization: There are no sanitization steps or validation rules defined for the external content before it is processed by the AI.
[SAFE]: The skill consists purely of documentation and advice. It does not include executable code, scripts, or requests for sensitive permissions. All external links and services mentioned (such as Toptal, Upwork, and GitHub) are well-known platforms referenced in a legitimate context.

hiring