The Agent Skills Directory

[DATA_EXFILTRATION]: The skill documentation includes hardcoded absolute file paths from the author's local environment.
Evidence: python C:\Users\renat\skills\agent-orchestrator\scripts\scan_registry.py in SKILL.md.
Impact: Exposes the directory structure of the developer's machine and leads to broken functionality in other environments.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from legal processes, which is a known vector for indirect prompt injection.
Ingestion points: User-provided auction edicts, court documents, and registration certificates as described in the usage scenarios.
Boundary markers: The instructions lack explicit delimiters or warnings to ignore embedded instructions within the processed legal text.
Capability inventory: The skill can execute local Python scripts (e.g., scripts/governance.py) and utilize various CLI tools listed in the frontmatter.
Sanitization: There is no evidence of sanitization or filtering logic for the input documents.
[COMMAND_EXECUTION]: The skill provides explicit instructions for the agent or user to execute shell commands using local scripts.
Evidence: Commands such as python agent-orchestrator/scripts/match_skills.py are provided for operation.
Description: While intended for functionality, these commands rely on the presence of an external 'agent-orchestrator' setup not fully contained within this skill.

leiloeiro-juridico