The Agent Skills Directory

[SAFE]: No malicious patterns, prompt injections, credentials exposure, or unauthorized remote connections were detected across any of the files.
[INDIRECT_PROMPT_INJECTION]: Evaluation of the untrusted data ingestion surface:
Ingestion points: File paths, directory hierarchies, and export configurations are ingested from the workspace via Read, Glob, and Grep tools as specified in SKILL.md.
Boundary markers: Absent. There are no explicit instructions or structural wrappers telling the model to ignore potential prompt instructions hidden within codebase strings or file names.
Capability inventory: The skill utilizes the Edit tool to modify file imports and the Bash tool to perform git mv commands as specified in SKILL.md.
Sanitization: Absent. The input is analyzed directly without pre-filtering.
Mitigation: The risk is low and inherent to the core purpose of code-quality auditing tools. The model is constrained to deterministic linting rules (such as kebab-case enforcement), minimizing the likelihood of execution from indirect prompt injection.

naming-format