The Agent Skills Directory

[SAFE]: The skill implements legitimate security utility functions. The threat_modeler.py script provides a local database of STRIDE/DREAD categories for component analysis, and secret_scanner.py identifies hardcoded credentials using transparent regex patterns. The skill's behavior matches its stated purpose as a security auditing tool.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted source code through its secret scanning and threat modeling workflows, which creates a potential surface for indirect prompt injection if an attacker-controlled file contains malicious instructions designed to mislead the agent. However, the tools are diagnostic and do not include automated actions based on the content.
Ingestion points: scripts/secret_scanner.py (reads content from user-specified project directories).
Boundary markers: None present for the scanned file content.
Capability inventory: scripts/secret_scanner.py (file-system read access), scripts/threat_modeler.py (CLI tool execution), and report output to the agent's context.
Sanitization: The scanner uses regex-based detection and masks identified secrets in its output, but it does not sanitize or filter the raw content of files being processed.

senior-security