The Agent Skills Directory

[DATA_EXFILTRATION]: The script scripts/report_violation.py contains functionality to send performance reports to an external Slack webhook URL using urllib.request. The target URL is provided by the user through command-line arguments. This constitutes a network operation to a non-whitelisted domain.
[INDIRECT_PROMPT_INJECTION]: The skill processes performance metrics and budget configurations from external JSON files to generate remediation recommendations. This ingestion of untrusted external data creates a potential surface for indirect prompt injection.
Ingestion points: The skill reads metrics from files such as metrics.json and lighthouse.json within scripts/validate_budget.py and scripts/report_violation.py.
Boundary markers: There are no explicit delimiters or specific instructions to the agent to ignore potentially embedded instructions in the ingested data.
Capability inventory: The skill has capabilities to read/write files and execute specific shell commands like lighthouse and webpack.
Sanitization: No sanitization or validation of string content is performed on the ingested metric data before the agent uses it to formulate recommendations.

validating-performance-budgets