The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external sources like audit findings, project plans, and existing documentation to generate strategic roadmaps. This creates a surface where malicious instructions hidden in those source materials could influence the agent's planning logic or output. \n
Ingestion points: The workflow in SKILL.md and references/roadmap-method.md explicitly instructs the agent to read docs/NORTH_STAR.md, existing plans, audit artifacts, and code state to assess gaps. \n
Boundary markers: There are no defined delimiters or specific instructions to isolate or ignore potential commands embedded within the documentation it processes. \n
Capability inventory: The skill performs file reading and writing operations within the project workspace, specifically updating docs/NORTH_STAR.md, roadmap files, and Architecture Decision Records (ADRs). \n
Sanitization: No sanitization or content validation mechanisms are described for the data retrieved from project files before it is processed by the agent.

north-star