The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions provide complex shell command pipelines for code analysis, utilizing tools such as grep, awk, and shell loops to process file contents.\n
Evidence: Step 1 and Step 2 in SKILL.md contain commands like grep -n "function\\|def\\|func " **/*.{js,ts,py,go} | while read line; do ... done and regex-based searches for SQL injection risks and hard-coded secrets.\n- [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it is designed to read and process the contents of untrusted external files.\n
Ingestion points: The skill uses the Read, Grep, and Glob tools to ingest data from various source files and datasets as specified in SKILL.md.\n
Boundary markers: There are no specific instructions or delimiters defined to prevent the agent from following malicious instructions that might be embedded within the files being scanned.\n
Capability inventory: The skill is limited to the Read, Grep, and Glob tools as per the allowed-tools metadata in SKILL.md, limiting potential impact of injection.\n
Sanitization: The skill does not implement explicit sanitization or filtering of the file contents before they are processed by the agent's logic.

pattern-detection