finding-deleted-feature-flags
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script (
scripts/strip_deleted_suffix.py) usingpython3to process feature flag data. The script is bundled within the skill and performs deterministic string manipulation on JSON inputs to recover original flag keys.\n- [DYNAMIC_EXECUTION]: The skill utilizes shell piping (echo ... | python3 ...) to run a provided Python script at runtime for data transformation. The script is static and included in the skill's distribution.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources, presenting a surface for indirect prompt injection.\n - Ingestion points: Data is retrieved from the
system.feature_flagstable (viaposthog:execute-sql) and flag activity logs (viaposthog:feature-flags-activity-retrieve), which contain user-generated strings like emails and names.\n - Boundary markers: The instructions do not specify the use of delimiters or safety warnings to separate system context from the retrieved activity log content in the final output.\n
- Capability inventory: The skill has access to SQL execution, activity log retrieval, and shell execution of local scripts.\n
- Sanitization: The provided Python script performs string suffix removal but does not implement validation or filtering to sanitize potential injection content within the processed data fields.
Audit Metadata