The Agent Skills Directory

[SAFE]: The skill provides instructions and reference patterns for code review. No malicious patterns, obfuscation, or unauthorized data access were detected.
[INDIRECT_PROMPT_INJECTION]: The skill analyzes user-provided source code via the Read, Glob, and Grep tools. While this constitutes an ingestion point for untrusted data, the skill lacks dangerous capabilities like network access or code execution that could be exploited via injection.
Ingestion points: User-specified files or search results from the local repository accessed via standard tools (SKILL.md).
Boundary markers: The skill instructions do not define explicit delimiters or instructions to ignore embedded content within the processed code.
Capability inventory: The skill is restricted to read-only tools (Read, Glob, Grep) and does not have write, network, or execution capabilities across any of its scripts or instructions.
Sanitization: No explicit sanitization or filtering of the ingested content is defined.
[DATA_EXFILTRATION]: The skill does not perform any network requests or attempt to exfiltrate data. It operates entirely within the local environment using standard read-only tools.
[CREDENTIALS_UNSAFE]: The reference documentation correctly identifies hardcoded secrets as a security risk and recommends using environment variables or the Keychain, demonstrating adherence to security best practices.

coding-best-practices