The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for consuming untrusted project data (requirements, source code, ADRs) to generate planning artifacts. While this establishes an attack surface for indirect prompt injection from those data sources, the skill itself lacks executable capabilities. * Ingestion points: SKILL.md indicates ingestion of external requirement and configuration files. * Boundary markers: Not defined in the text. * Capability inventory: No command execution or network tools are present in the skill files. * Sanitization: None defined.
[NO_CODE]: There are no script files (e.g., .py, .js, .sh) or binaries in the skill package, which precludes common attack vectors like remote code execution or malware persistence.

dl-plan