The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides shell commands to read and modify local configuration files to switch between model profiles.
Evidence: Use of sed -i.bak, awk, and grep to manage .abilities/settings.yaml. These operations are consistent with the skill's stated purpose of configuration management.
[DATA_EXFILTRATION]: The skill accesses local settings files to determine the active model profile.
Evidence: Accesses .abilities/settings.yaml and ~/.claude/abilities/settings.yaml. There is no evidence of sensitive data exposure or network exfiltration.
[PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by processing local configuration data.
Ingestion points: Reads configuration from .abilities/settings.yaml.
Boundary markers: No explicit boundary markers or delimiters used for configuration parsing.
Capability inventory: Provides commands for file modification (sed) and reading (grep, awk).
Sanitization: No input validation or sanitization is performed on the configuration values.

model-profiles