The Agent Skills Directory

[PROMPT_INJECTION]: The skill functions as a 'skill chameleon' that extracts and applies patterns from other plugins. This design creates a surface for indirect prompt injection where malicious instructions hidden in a plugin's documentation or code could influence the agent's behavior.
Ingestion points: Extracts capability patterns from plugins/community/, plugins/packages/, and plugins/examples/ (referenced in SKILL.md).
Boundary markers: Absent; there are no instructions to use delimiters or ignore embedded instructions within the analyzed plugins.
Capability inventory: The skill utilizes Bash(cmd:*), Read, Grep, and Glob (referenced in SKILL.md).
Sanitization: Absent; the skill does not perform validation or sanitization on the content extracted from external plugins before applying it to the current task.
[COMMAND_EXECUTION]: The skill requests unrestricted bash access (Bash(cmd:*)) to facilitate plugin discovery and analysis using tools like find and grep. While used for legitimate functionality in the provided scripts (e.g., plugin_analyzer.py), the broad scope of the permission increases the potential impact of any command injection or malicious influence from external plugins.

skill-adapter