The Agent Skills Directory

[PERSISTENCE]: The skill configures a persistent OpenClaw hook and an optional cron job to provide continuous background security monitoring. These mechanisms are documented and include preflight reviews to ensure the user is aware of the automation before it is enabled.
[EXTERNAL_DOWNLOADS]: The skill fetches security advisory feeds, catalog metadata, and updates from the vendor's official infrastructure (prompt.security) and GitHub releases. These operations are essential for threat intelligence and are protected by mandatory cryptographic signature verification.
[INDIRECT_PROMPT_INJECTION]: The skill processes external advisory data and displays it to the agent/user. It mitigates potential injection risks by using structured data formats (JSON), clear notification headers, and requiring explicit human approval for any recommended security actions or skill removals.
[COMMAND_EXECUTION]: Uses standard system utilities (curl, jq, openssl, unzip) and the openclaw CLI to manage the security environment. Command usage is scoped to the skill's management functions and does not involve executing unsanitized user or network input.

clawsec-suite