The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes Python scripts (sample-size-calculator.py and significance-tester.py) located within the plugin's root directory via shell commands.
Evidence: Step 10 of the process executes python "${CLAUDE_PLUGIN_ROOT}/scripts/sample-size-calculator.py" and python "${CLAUDE_PLUGIN_ROOT}/scripts/significance-tester.py" with interpolated arguments such as {stage-rate}, {mde}, and {n}.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from various local files which could be influenced by external sources or third-party configurations.
Ingestion points: The skill reads brand profiles (profile.json), active brand markers (_active-brand.json), compliance rules (compliance-rules.md), guidelines (_manifest.json), and agency SOPs from the ~/.claude-marketing/ directory.
Boundary markers: No explicit boundary markers or instructions to ignore embedded instructions are defined for the imported content.
Capability inventory: The skill has the capability to execute shell commands via Python scripts and read/write access to the brand's local directory.
Sanitization: There is no evidence of sanitization or validation for the data interpolated into the prompt or the shell command arguments.

funnel-audit