bash-linux-v2
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [COMMAND_EXECUTION]: Provides functional patterns for process management and network investigation, including tools for terminating processes (
kill) and identifying network listeners (lsof). - [EXTERNAL_DOWNLOADS]: Explicitly references a community GitHub repository (
github.com/sickn33/antigravity-awesome-skills) as the upstream source for the skill's content and provenance documentation. - [INDIRECT_PROMPT_INJECTION]: Features a vulnerability surface where user-provided task descriptions are interpolated into Bash command templates, potentially allowing for command injection if user input is not sanitized.
- Ingestion points: User tasks defined in skill examples.
- Boundary markers: Absent; the skill does not employ delimiters to isolate user-supplied data from command logic.
- Capability inventory: Includes templates for file operations, process control, and network interactions using standard Linux utilities.
- Sanitization: None observed in the pattern documentation.
- [METADATA_POISONING]: Identified a discrepancy between the author listed in the skill metadata (
sickn33) and the platform-provided author context (diegosouzapw).
Audit Metadata