The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions designed to override agent safety protocols or ignore system instructions were found. The skill includes explicit safety boundaries for mental health and legal issues.- [Data Exposure & Exfiltration] (SAFE): No sensitive file paths, hardcoded credentials, or network operations are present. The skill proactively advises users to use anonymized summaries and avoid sharing secrets.- [Obfuscation] (SAFE): No encoded content, zero-width characters, or homoglyphs were detected across the skill files.- [Unverifiable Dependencies & Remote Code Execution] (SAFE): The skill does not define any external package dependencies or perform remote script execution.- [Privilege Escalation] (SAFE): No commands related to privilege escalation (e.g., sudo, chmod) were found.- [Indirect Prompt Injection] (SAFE): The skill ingests untrusted user data (facts/examples) as documented in SKILL.md and references/INTAKE.md. It provides boundary markers via the 'Stop' checks in references/CHECKLISTS.md. However, since the skill has zero executable capabilities (no-code), the risk of exploitation is non-existent.

having-difficult-conversations