The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local shell commands to audit the host system and manage services.
It executes diagnostic commands including ps, ss, systemctl, launchctl, and docker to inspect running processes and network listeners (SKILL.md).
It utilizes internal management scripts such as scripts/openclaw-auth-profile-sync.sh and scripts/normalize-openclaw-models.py for bot configuration tasks.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by processing external system data.
Data ingestion points include host logs and command outputs from system utilities, which may contain untrusted content (SKILL.md).
The prompt construction logic does not define explicit boundary markers or delimiters for the interpolation of system outputs.
[DATA_EXFILTRATION]: The skill retrieves system metadata but implements privacy guardrails.
The auditing workflow identifies Unix users, service accounts, and system directory structures to build an operational map (SKILL.md).
Guardrails: The instructions include strict documentation rules mandating the redaction of passwords, tokens, IP addresses, and identifying user details before any output is generated (SKILL.md).
[EXTERNAL_DOWNLOADS]: The skill fetches documentation and references official installation utilities from external sources.
It downloads a local mirror of documentation from https://docs.openclaw.ai using the sync-openclaw-docs command (SKILL.md).
The documentation provided with the skill includes instructions for installing tools using shell scripts from openclaw.ai, bun.sh, and github.com (references/openclaw-docs/current/).

server-doctor