The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project statistics and workflow data from a file located at $HOME/.claude/get-shit-done/workflows/stats.md to guide its execution.
Ingestion points: The agent ingests context from the execution_context tag referencing the stats.md file.
Boundary markers: The instructions lack explicit delimiters or directives to ignore embedded instructions within the ingested workflow file.
Capability inventory: The skill is authorized to use Bash and Read tools, which could be exploited if the external file contains malicious directives.
Sanitization: There is no evidence of sanitization, filtering, or validation for the content read from the file before the agent processes it.

gsd-stats