The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform repository analysis and metadata extraction.
Executes ls and grep on files in the .github/workflows/ directory to identify active component IDs.
Executes git diff to extract changes from the current pull request for review against the base branch.
[DATA_EXFILTRATION]: The skill retrieves telemetry and usage data from the vendor's internal database.
Queries the KBC_USE4_37 Snowflake database using mcp__keboola__* tools to obtain job statistics and configuration details.
Accesses potentially sensitive configuration JSON fields to determine parameter usage.
Includes specific safety instructions to anonymize all telemetry data and strictly forbids the inclusion of client or project identifiers in public output.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes untrusted data from the repository being reviewed.
Ingestion points: Data is ingested from workflow files (push*.yml) and code diffs across multiple directories.
Boundary markers: No specific boundary markers or 'ignore' instructions are used when processing the extracted text or diffs.
Capability inventory: The agent can execute shell commands via Bash and run SQL queries via MCP tools.
Sanitization: The skill lacks explicit sanitization or validation for component IDs extracted from repository files before they are interpolated into SQL query strings, which could theoretically be manipulated to influence query logic.

backward-compatibility-reviewer