leiloeiro-edital
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is intended to ingest and analyze external auction notices which are untrusted data sources.
- Ingestion points: The "Protocolo De Análise De Edital" in SKILL.md describes how the agent should process received notice information.
- Boundary markers: Absent. There are no specific instructions or delimiters provided to prevent the agent from following instructions potentially embedded within the auction notices.
- Capability inventory: The file scripts/governance.py implements file-writing capabilities to a local log file (action_log.jsonl).
- Sanitization: Absent. No logic is present to sanitize or escape data before it is processed by the agent's persona.
- [COMMAND_EXECUTION]: The skill documentation provides command-line examples that reference absolute paths specific to the author's local environment.
- Evidence: SKILL.md contains a command for registry scanning that points to 'C:\Users\renat\skills\agent-orchestrator\scripts\scan_registry.py'. While these are manual instructions, referencing local developer paths is an insecure documentation practice.
Audit Metadata