The Agent Skills Directory

[COMMAND_EXECUTION] (MEDIUM): The skill calls a local script ./scripts/log-skill.sh with the current PID. Since the script source is not provided, its actions cannot be fully verified.
[PROMPT_INJECTION] (MEDIUM): The skill identifies as having a surface for indirect prompt injection. 1. Ingestion points: Reads contents of files within docs/solutions/ using grep and ls. 2. Boundary markers: Absent; the skill does not wrap processed content in delimiters. 3. Capability inventory: Includes file modification and execution of the logging script. 4. Sanitization: Absent; the agent is instructed to format and promote found patterns into critical documentation, which could lead to persistent malicious instructions in the solution database.

compound-docs