The Agent Skills Directory

[SAFE]: The skill's functions and instructions are entirely consistent with its purpose as a pre-change impact analyst. It does not contain any suspicious code, obfuscation, or unauthorized network operations.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted code and change descriptions, which represents a potential surface for indirect prompt injection, although no malicious use was detected.
Ingestion points: Target source files and change descriptions are processed during the SCOPE and VERTICAL analysis phases.
Boundary markers: The instructions do not explicitly mandate delimiters or safety warnings for the content being analyzed.
Capability inventory: The skill utilizes shell-based analysis tools including grep, find, and madge for dependency tracing.
Sanitization: No explicit sanitization or input validation strategies for the ingested content are mentioned.
[PROMPT_INJECTION]: The skill uses detailed instructional language and references external engineering policies to guide the agent's decision-making process.
Evidence: The skill references 'Amazon 2026' and 'OWASP 2026 Agentic Top 10' policies to set thresholds for risk scoring and impact assessment.

ripple