The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted markdown data from several files, presenting a surface for indirect prompt injection where data content could potentially influence the agent's logic.
Ingestion points: Reads content from spec.md, decisions.md, and plan.md files, as well as an external artifact path specified in the Source: field of the plan.
Boundary markers: The instructions do not specify any markers or delimiters to isolate processed data from the agent's instructions.
Capability inventory: The skill has the capability to read project files and write modified plan files, as well as call the /emit-pending-decisions tool.
Sanitization: The skill processes and edits markdown content without explicit sanitization or validation of the input data.

reconcile-plan