The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill utilizes a shell command in the frontmatter (!ls production/sprints/) to provide the agent with a list of existing files. This is a legitimate use of the platform's dynamic context feature to establish project-level awareness at load time.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from various untrusted sources including design documents (design/gdd/) and risk registers. While these files could theoretically contain malicious instructions, the skill's activities are restricted to gathering context for report generation, which is a standard and low-risk operation.
Ingestion points: Files in production/milestones/, production/sprints/, design/gdd/, and production/risk-register/.
Boundary markers: None explicitly defined for the ingested data.
Capability inventory: File reading, file writing (production/sprints/, production/sprint-status.yaml), and spawning tasks via the Task tool.
Sanitization: None specified for the content read from external files.
[COMMAND_EXECUTION]: The skill uses the Task tool to spawn a feasibility gate (PR-SPRINT). This is used for internal workflow validation (producer review) and does not involve arbitrary command execution or system-level risk.

sprint-plan