The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses Bash to execute local Python scripts that perform the analysis. These scripts accept user-provided task descriptions and session IDs as command-line arguments. While the Python implementation uses the safe subprocess.run method with argument lists to avoid shell injection, the initial invocation in SKILL.md relies on environment variable substitution which depends on platform-level sanitization.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because its core function is to read and summarize content from the repository's source code files.
Ingestion points: The scripts analyze.py and analyze_codebase.py recursively scan and read the contents of source files across the entire repository to calculate similarity and extract patterns.
Boundary markers: The output report does not include explicit boundary markers or instructions to the agent to ignore potentially malicious commands embedded within the code comments or documentation it summarizes.
Capability inventory: The agent invoking this skill has access to the Bash and Read tools, providing a path for executing instructions if it is tricked into following a malicious pattern extracted from the codebase.
Sanitization: The skill extracts keywords and function signatures using regular expressions but does not sanitize the text for prompt injection patterns before presenting implementation suggestions to the agent.

codebase-analysis