The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes qdbus to execute commands against the org.kde.kwalletd6 service to open, read from, and close the system wallet.
[DATA_EXFILTRATION]: The skill is designed to retrieve sensitive system-level secrets, including passwords and binary keys (e.g., 'github-token'), from the KWallet vault. This programmatic exposure of credentials to the AI agent's execution context is a high-risk activity.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves data from a system service that could contain malicious instructions if a wallet entry is compromised.
Ingestion points: readPassword, readMap, and readEntry calls in SKILL.md.
Boundary markers: None; the skill does not include delimiters or instructions to treat the retrieved content as untrusted data.
Capability inventory: Command execution via qdbus and underlying shell access.
Sanitization: No sanitization or validation of the retrieved secret values is performed before they enter the agent's prompt context.

kwallet-read