swfactory
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external 'intent/spec/plan artifacts' which creates a surface for instructions embedded in data to influence agent behavior.\n
- Ingestion points: The agent is instructed to follow 'intent/spec/plan artifacts' when running in 'Inner stage mode'.\n
- Boundary markers: The instructions do not define specific delimiters or warnings to separate data from instructions within these artifacts.\n
- Capability inventory: The skill utilizes the
swfCLI for job submission and status monitoring, and mentions execution within sandboxes.\n - Sanitization: There are no explicit mentions of sanitizing or validating the contents of the artifacts before processing.\n- [EXTERNAL_DOWNLOADS]: The skill provides instructions to install a Node.js package from a remote source.\n
- Package:
zozo123/ariflow-swfactoryinstalled vianpx.\n - Context: The package is provided by the skill author and facilitates the CLI integration.\n- [COMMAND_EXECUTION]: The skill utilizes command-line tools to manage the software factory lifecycle.\n
- Patterns: Usage of
swf(doctor, submit, list) and a wrapper scriptscripts/swf_harness.sh.
Audit Metadata