The Agent Skills Directory

[SAFE]: The skill is a standard development tool for reviewing other skills and does not exhibit any malicious patterns such as data exfiltration, obfuscation, or unauthorized access.
[COMMAND_EXECUTION]: The skill executes a bundled script, scripts/skill-lint.sh, to perform deterministic checks on skill directories. The script uses safe text-processing utilities like awk, grep, and find to analyze file structure and metadata without executing the contents of the analyzed files.
[PROMPT_INJECTION]: The skill processes untrusted data (other skill files) as part of its review process, creating a surface for indirect prompt injection. This risk is assessed as safe because the capability is necessary for the skill's primary purpose and the analysis is restricted to qualitative review.
Ingestion points: Reads SKILL.md and referenced files from user-specified directories.
Boundary markers: None explicitly defined for the input content.
Capability inventory: Executes local bash scripts, reads file system contents, and generates text reports.
Sanitization: No specific sanitization or filtering of input content was observed.

skill-improver