The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): The skill documentation uses generic placeholders (e.g., 'admin@example.com', 'TASK001') for examples. No hardcoded credentials, API keys, or sensitive environment paths were detected.
[Prompt Injection] (SAFE): No instructions attempting to override agent behavior, bypass safety filters, or extract system prompts were found. The use of the word 'RULE' and 'IMPORTANT' is contextually appropriate for developer documentation.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): There are no commands for installing external packages or executing remote scripts. The provided code snippets are intended for use within an existing Frappe/ERPNext environment.
[Obfuscation] (SAFE): No Base64, zero-width characters, or homoglyphs were detected. The content is plain-text markdown and JSON.
[Metadata Poisoning] (SAFE): The metadata fields (name, description, author) accurately reflect the content of the skill. The presence of GitHub 404 error responses in several reference files (e.g., references/methods.md) appears to be a benign technical artifact from the skill's creation process and contains no malicious instructions.
[Privilege Escalation] (SAFE): While the documentation mentions ignore_permissions=True, it does so in the context of teaching developers how to use Frappe API flags and includes warnings about the associated risks (anti-patterns).

erpnext-database