The Agent Skills Directory

[SAFE]: The skill provides instructional guidelines for generating Salesforce Custom Setting metadata and associated CLI commands.
[SAFE]: Includes explicit security warnings in Section 5, correctly identifying that Custom Settings are not suitable for storing secrets (API keys, tokens, etc.) and recommending the use of Named Credentials instead.
[SAFE]: Implements a 'no-silent-downgrade' security policy for component visibility, ensuring that the agent does not automatically lower security settings (e.g., from Protected to Public) if a deployment fails.
[INDIRECT_PROMPT_INJECTION]: The skill generates shell commands (sf data create record) using values provided by the user. This represents a potential injection surface; however, the skill mitigates this by providing specific instructions to the agent on how to properly quote and escape these values (Section 6).
Ingestion points: User-provided data for custom setting field values in SKILL.md.
Boundary markers: None explicitly defined for input, but generated commands are intended for inline display in chat responses.
Capability inventory: Shell command generation for the sf CLI.
Sanitization: Explicit instructions are provided to wrap the --values argument in double quotes and to single-quote individual values containing spaces or special characters.

platform-custom-setting-generate