The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external billing CSV exports, which could be used as a vector for indirect prompt injection if the data contains hidden instructions intended for the agent.
Ingestion points: The skill reads source files provided by the user in Step 3 and Step 5 (--in <source file>).
Boundary markers: There are no explicit instructions to use delimiters or warnings for the agent to ignore embedded instructions within the CSV content.
Capability inventory: The skill has the ability to execute shell commands, run Python scripts, and write files to the local filesystem (Step 5, Step 8).
Sanitization: The instructions do not specify any sanitization, escaping, or validation of the external content before it is processed by the agent or script.
[COMMAND_EXECUTION]: The skill uses shell commands to manage its internal files, check for configurations, and execute its core processing logic.
Evidence: Step 1 uses mkdir and rsync to set up templates from ${CLAUDE_PLUGIN_ROOT}. Step 5 and Step 6 execute python3 scripts/focus_export.py with various arguments. The Overlay section executes a shell check: [ -n "$POWERSUIT_OVERLAY" ] && ls -1 "$POWERSUIT_OVERLAY/finops" 2>/dev/null || true.

focus-export