The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute specific account management tools, including claude-acct for macOS and caam for other environments, to perform credential rotation.
[CREDENTIALS_UNSAFE]: The primary purpose of the skill is to manage and rotate authentication tokens and credential files. While no exfiltration or hardcoded secrets were detected, the skill operates directly on sensitive identity data.
[PROMPT_INJECTION]: The skill ingests untrusted data in the form of a 'caller-selected' account name, which presents a surface for indirect prompt injection as there are no specified boundary markers or sanitization steps before the input is passed to system tools.
Ingestion point: 'caller-selected coding-agent account' defined in SKILL.md.
Boundary markers: None detected.
Capability inventory: Execution of claude-acct and caam binaries.
Sanitization: No sanitization or validation of the account string is documented.

account-rotation