The Agent Skills Directory

[SAFE]: The skill implements a structured audit methodology using a restricted set of tools (Read, Grep, Glob) and does not perform network operations or execute arbitrary shell commands.
[SAFE]: No evidence of credential exposure, obfuscation, or persistence mechanisms was found in the skill instructions or metadata.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting and processing untrusted data from external firewall configuration files (e.g., .tf, .acl, iptables).
Ingestion points: The skill uses Glob and Grep to find and read various network configuration and Infrastructure-as-Code files.
Boundary markers: While the skill lacks explicit technical delimiters for data interpolation, it includes a 'Prompt Injection Safety Notice' advising the agent to treat configuration comments as untrusted data.
Capability inventory: The skill's capabilities are limited to reading files and generating a report; it lacks the ability to write to the filesystem or make external network requests, significantly limiting the potential impact of an injection.
Sanitization: The skill relies on natural language instructions for the agent to ignore instructions embedded in comments rather than programmatic sanitization.

firewall-review