The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The Python scripts process user-provided data, such as system names and team statistics, which are then formatted into reports. Because these inputs are not sanitized before being returned as text to the agent, they could potentially be used to host malicious instructions intended to influence the agent's behavior during the interpretation phase.
Ingestion points: The analyze_system method in scripts/tech_debt_analyzer.py (ingests system_data) and the calculate_scaling_plan method in scripts/team_scaling_calculator.py (ingests current_state).
Boundary markers: No explicit delimiters or instructions are used in the script output to help the agent distinguish between tool-generated reports and data fields that may contain untrusted text.
Capability inventory: Both scripts are strictly computational and lack dangerous capabilities; they do not perform network requests, file writes, or command execution.
Sanitization: The scripts do not escape or validate string inputs (e.g., system_name) before including them in the final output string.
[SAFE]: The provided Python scripts (tech_debt_analyzer.py and team_scaling_calculator.py) perform legitimate mathematical calculations for engineering metrics and technical debt scoring. The documentation and templates in the references/ directory follow industry-standard frameworks (DORA, ADR, SPACE) and present no security risk.

cto-advisor