The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for running a local Python script plugins/abstract/scripts/skill_graph.py and a test suite using pytest. These are standard operations for a development and auditing tool and are performed on local files within the project repository.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data by parsing Skill() references within other skill files to build a directed graph. This is the core functionality and is used for metadata analysis rather than execution of instructions found in external content.
[SAFE]: No security threats were identified across all analyzed categories, including prompt injection, data exfiltration, and obfuscation. The skill's behavior is consistent with its stated purpose of auditing skill compositions.

skill-graph-audit