The Agent Skills Directory

[SAFE]: No security issues detected. The skill provides legitimate utility for text transformation using the standard sed tool and refers to official documentation from trusted sources.
[PRIVILEGE_ESCALATION]: The installation scripts (scripts/install.ps1) and troubleshooting documentation (references/troubleshooting.md) mention the use of sudo for package installation and editing system-restricted files. This is standard administrative behavior for the intended use case and does not represent a malicious attempt at privilege escalation.
[INDIRECT_PROMPT_INJECTION]: The skill operates on text files which may contain untrusted data, presenting a potential surface for indirect prompt injection if an agent blindly follows instructions found in processed files.
Ingestion points: File paths and piped input provided to sed commands across all reference files and SKILL.md.
Boundary markers: None explicitly used in command templates.
Capability inventory: The skill utilizes sed for file reading and writing (via the -i flag or shell redirection) as documented in references/cheatsheet.md and references/advanced-usage.md.
Sanitization: None; the tool performs deterministic regex-based transformations as its primary function, though the documentation correctly suggests testing on sample data first.

sed