using-pb-spec
Pass
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill uses strong, mandatory language (e.g., "not negotiable", "ABSOLUTELY MUST") to ensure the agent prioritizes its workflow and tool usage over default system behaviors. While aggressive, it explicitly maintains that user-provided instructions in local configuration files (like CLAUDE.md or AGENTS.md) take the highest precedence, which serves as a safety guardrail.
- [PROMPT_INJECTION]: The skill defines an explicit instruction priority hierarchy, placing its own rules and user-defined files above the default system prompt. This is a standard steering mechanism for complex agent workflows but is noted as an override attempt.
- [DATA_EXFILTRATION]: The skill mentions reading configuration files like
AGENTS.mdandCLAUDE.md. While these are ingestion points for potentially untrusted data, the skill does not include any capabilities for exfiltrating this data to external servers or exposing credentials. - [REMOTE_CODE_EXECUTION]: No evidence of unauthorized package installation or remote script execution was found. The skill references local tool names and a structured workflow for development tasks.
Audit Metadata