The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructions for "Task" tickets advise the agent to record findings in the issue tracker, specifically mentioning "where credentials are kept" (onde ficam as credenciais). This encourages the exposure of sensitive security metadata in a shared environment.
[DATA_EXFILTRATION]: The skill is designed to move project information, investigation results, and "ideas" to an external issue tracker configured via the /setup-leandrocfe-skills command. This creates a channel where sensitive local code or documentation could be exfiltrated to an external service.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection. It retrieves and processes ticket content ("mapa", "corpo completo de qualquer ticket") from an external issue tracker. Maliciously crafted tickets in the tracker could influence the agent's behavior or decision-making process.
Ingestion points: External issue tracker (tickets and map issues).
Boundary markers: Absent; no instructions to ignore or delimit instructions found within the tickets.
Capability inventory: Access to external setup tools (/setup-leandrocfe-skills), decision-making logic, and the ability to post comments/resolutions to a tracker.
Sanitization: Absent; content from the tracker is used directly to resolve decisions.
[COMMAND_EXECUTION]: The skill utilizes a vendor-provided setup command /setup-leandrocfe-skills to configure the issue tracker environment. While this is a vendor-owned resource, it represents an external script execution dependency required for the skill's primary functionality.

wayfinder