The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied text to perform stylistic transformations ('humanization'). This creates an attack surface where untrusted data provided by the user could potentially contain instructions aimed at influencing the agent's behavior during the rewriting process.
Ingestion points: User-provided drafts or text samples intended for humanization, as identified in SKILL.md.
Boundary markers: The instructions do not define explicit delimiters or use 'ignore embedded instructions' warnings to separate the input text from the rewriting task.
Capability inventory: The skill utilizes the @rephrasy/mcp humanize tool and provides specific linguistic heuristics for the agent to follow.
Sanitization: No sanitization, escaping, or validation mechanisms for the external content are specified.
[EXTERNAL_DOWNLOADS]: The skill references the @rephrasy/mcp package available via NPM. This is a vendor-provided resource from 'rephrasyai' intended to facilitate the humanization of text.

bypass-zerogpt