The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access detected. The skill instructions are standard for software development workflows.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it is designed to ingest and act upon instructions from external documentation files (PRODUCT.md, TECH.md). This risk is inherent to its primary purpose of spec-driven development.
Ingestion points: Reads specifications from PRODUCT.md and TECH.md as described in SKILL.md (Workflow Step 1).
Boundary markers: No explicit delimiters or instructions to ignore potentially malicious embedded commands are present.
Capability inventory: The agent is tasked with file creation (e.g., PROJECT_LOG.md), code implementation, and running verification tests as part of the implementation workflow.
Sanitization: No sanitization or validation of the content within the input documentation files is specified.

implement-specs