The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing dependency skills from the author's GitHub repository (https://github.com/konglong87/superPM) using a CLI command (npx skills add). As these resources are owned by the skill author ('konglong87'), this represents expected functionality for a multi-skill workflow architecture.
[INDIRECT_PROMPT_INJECTION]: The workflow orchestrates tools that process external information such as market trends and competitor data. This presents a potential surface for indirect injection if sub-skills ingest untrusted data without proper sanitization.
Ingestion points: Data processed by the dependency skills pm-market, pm-competitor, and pm-demand (e.g., via WebSearch).
Boundary markers: None explicitly defined in the orchestration instructions to delimit untrusted content.
Capability inventory: The skill is allowed to use Bash, Write, Read, and WebSearch tools.
Sanitization: Not explicitly implemented at the orchestration level; sub-skills are responsible for their own data handling.

discover