The Agent Skills Directory

[SAFE]: The skill implements a robust security policy for all outgoing notifications.
Evidence: The skill documentation enforces a 'Sequential sending confirmation
no exceptions' (逐次发送确认——无例外) rule, requiring the agent to display message content and recipient and wait for explicit consent before using enterprise communication tools.
[PROMPT_INJECTION]: The skill possesses a potential surface for indirect prompt injection through the ingestion of external tracking data, though this is mitigated by the mandatory human review step.
Ingestion points: Data is read from ~/.claude/plugins/config/claude-for-legal/regulatory-legal/gap-tracker.yaml and references/comment-tracker.yaml, which are populated by other upstream skills.
Boundary markers: No explicit delimiters or 'ignore' instructions are specified for the interpolation of data fields into reports or notifications.
Capability inventory: The skill can format and send messages to enterprise communication platforms (e.g., Slack) via the owner_contact and owner_slack fields.
Sanitization: No specific sanitization or validation of the ingested data is described; the system relies on the mandatory user review checkpoint to identify and block malicious content before transmission.

gap-surfacer