The Agent Skills Directory

[PROMPT_INJECTION]: No malicious override or bypass instructions were detected. The instructions follow standard task-oriented guidance.
[DATA_EXFILTRATION]: No evidence of sensitive file access or unauthorized data transmission. The skill uses placeholders for user links and explicitly instructs the model not to invent URLs.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns or unverified dependencies found.
[COMMAND_EXECUTION]: The skill does not execute shell commands or use dynamic execution patterns like eval or exec.
[OBFUSCATION]: No hidden or encoded content (Base64, hex, homoglyphs, or zero-width characters) was detected in the markdown or reference files.
[INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface through WebSearch and WebFetch (SKILL.md Step 2). While it processes external data, the instructions are scoped to extracting specific patterns and keywords for description generation, representing a standard low-risk capability.

create-description