The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface because it is designed to ingest and process untrusted code from external files. 1. Ingestion points: Code files provided to the agent for auditing (as described in the methodology). 2. Boundary markers: Absent; the instructions do not specify the use of delimiters or warnings to ignore instructions embedded within the code being audited. 3. Capability inventory: The skill requires reading and writing files and executing shell commands to run linters and formatters. 4. Sanitization: Absent; there is no explicit instruction to sanitize or escape the content of the files being processed. This surface is inherent to the task of code auditing and does not represent a malicious instruction within the skill itself.

style-audit