The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process arbitrary text provided by the user to identify and rewrite AI writing patterns. Because the skill has access to powerful file system modification tools and lacks explicit instructions to treat input text as non-executable data or to use secure boundary markers, it is susceptible to indirect prompt injection. A malicious user could provide text containing hidden instructions that the agent might execute rather than humanize.
Ingestion points: Processes user-supplied text during the humanization task as described in the 'Your Task' section of SKILL.md.
Boundary markers: Absent; the skill does not define specific delimiters (like XML tags) or instructions to ignore embedded commands within the input text.
Capability inventory: Access to Read, Write, Edit, Grep, Glob, and AskUserQuestion tools as defined in SKILL.md.
Sanitization: Absent; no logic is present to filter, escape, or validate instructions found within the input text.
[EXTERNAL_DOWNLOADS]: The skill's documentation and metadata point to a third-party GitHub repository for source code and installation.
Source: https://github.com/blader/humanizer referenced in both README.md and SKILL.md.
Context: The README.md provides git clone instructions for installation, and SKILL.md credits this repository as the original source of the skill.

humanizer