The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes utilities that ingest and process external data sources, creating a potential surface for indirect injection attacks.
Ingestion points: scripts/audit_log_analyzer.py reads external Vault audit log files via Path.read_text, and scripts/rotation_planner.py reads JSON inventory files.
Boundary markers: The skill does not implement explicit boundary markers or instructions for the agent to ignore embedded commands within the ingested log/inventory data.
Capability inventory: The skill possesses file-writing capabilities via scripts/vault_config_generator.py and execution context for Python scripts.
Sanitization: Data parsed from logs (such as metadata or error messages) is interpolated into reports without extensive sanitization against control characters or prompt instructions.
[DYNAMIC_EXECUTION]: The scripts/vault_config_generator.py script dynamically generates shell scripts (setup_*.sh) and HCL policy files based on user-provided flags.
Artifact Generation: It constructs executable setup scripts for various Vault secrets engines (KV, PKI, AWS, etc.) and auth methods.
Review Requirement: While these scripts are intended for manual review and application by a systems administrator, they represent the generation of executable content at runtime.

secrets-vault-manager