The Agent Skills Directory

[COMMAND_EXECUTION]: Employs standard developer tools like git, grep, find, and tail to investigate codebase state and log files. It also runs a bundled local script scripts/debug_report.py to create report templates.
[DATA_EXFILTRATION]: Accesses environment variables filtered for the string 'debug' to help identify configuration-related issues. This operation is constrained and aligns with the primary purpose of the skill.
[PROMPT_INJECTION]: Ingests potentially untrusted data from log files and error messages, creating a surface for indirect prompt injection. Ingestion points: logs/error.log and user error strings. Boundary markers: None. Capability inventory: Bash, Write, Edit, Read, Grep, Glob tools are available to the agent. Sanitization: No explicit sanitization or filtering of log content is performed. This surface is considered low risk and inherent to the skill's primary function of log analysis.

debugger