The Agent Skills Directory

[SAFE]: No malicious behavior or security concerns identified. The skill correctly identifies its boundaries by routing implementation and debugging tasks to specialized skills. All external references point to official and well-known documentation sources (e.g., Martin Fowler, Google, Playwright).
[PROMPT_INJECTION]: The skill evaluates untrusted user input provided in 'policy packets' to generate validation briefs. While it lacks explicit sanitization or boundary markers (delimiters) for this data ingestion point, the risk is mitigated by the skill's instructions which prioritize normalizing the data into a structured brief rather than executing the content of the input. Capability inventory includes the Bash, Read, and Write tools as defined in the frontmatter, but no instructions trigger dangerous use of these capabilities on untrusted data.

testing-strategies