The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill requires administrative privileges to perform debugging tasks. It uses sudo for multiple high-impact operations, including installing system packages (apt install auditd), configuring kernel-level auditing (auditctl), and managing system services and timers via systemctl. These actions allow the agent to modify the host system's security posture and service configuration.
[CREDENTIALS_UNSAFE]: The provided diagnostic commands instruct the agent to monitor and record changes to ~/.hermes/.env. One specific command appends the entire contents of this file to a plaintext log (~/env_changes.log). Environment files commonly contain sensitive data like API keys, proxy credentials, and authentication tokens; duplicating this data into a non-standard log file increases the risk of credential exposure.
[COMMAND_EXECUTION]: The skill executes various shell commands to inspect the environment, including process analysis (ps aux), cron job inspection (crontab), and scanning shell configuration files (.bashrc, .zshrc, etc.) for specific patterns. It also utilizes specialized CLI tools (hermes, cordon) to modify application states.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and processes untrusted data from local configuration and profile files.
Ingestion points: Reads and greps content from user-controlled files such as ~/.hermes/.env, ~/.bashrc, ~/.zshrc, and /etc/cron.d/.
Boundary markers: None identified; the skill processes file contents directly using standard shell utilities without using delimiters or instructions to ignore embedded commands.
Capability inventory: Significant system capabilities including sudo access, package installation, and the ability to modify system services and user shell configurations.
Sanitization: There is no evidence of validation or sanitization of the content retrieved from these files before it is processed or used in further commands.

env-auto-recover-debug