The Agent Skills Directory

Indirect Prompt Injection (HIGH): The skill is designed to ingest and analyze external data (code and documentation) which is inherently untrusted.
Ingestion points: Processes user-provided codebases or documentation files via Read and Grep tools.
Capability inventory: Includes Bash, Write, and Edit tools, allowing for arbitrary command execution and file system modification across the environment.
Boundary markers: Absent. The skill instructions do not define delimiters to separate untrusted data from instructions, increasing the likelihood of the model obeying embedded commands.
Sanitization: Absent. There is no mention of filtering or validating the content of the documentation being analyzed before processing.
Risk: An attacker could embed malicious instructions within a README or source code comment that, when processed by this skill, triggers harmful bash commands or unauthorized file edits.

code-documentation-analyzer